Technology in Architectural Practice

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

Artificial Intelligence has actually been with us for some time in various early forms, including expert systems and even simple programming. Technology has been far more helpful than a hindrance, and I am sure each of us could cite many positive outcomes. In many forms, I hear that our creativity is at risk from the use of codes, standards, and even tools to support our ability to do our jobs. Our challenges are far greater than worrying about AI negatively. We still have a long way to go before it replaces creativity. On this day after Frank Gehry's passing, I note that he embraced technology as it actually helped him to be more creative. Let's do everything in our power to ensure it is used as a tool to make us better, and not let it be in charge. With every new capability we discover, some people will use it positively, while others will abuse it. It takes a far more creative person to use it to their advantage. 

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

Artificial Intelligence has actually been with us for some time in various early forms, including expert systems and even simple programming. Technology has been far more helpful than a hindrance, and I am sure each of us could cite many positive outcomes. In many forms, I hear that our creativity is at risk from the use of codes, standards, and even tools to support our ability to do our jobs. Our challenges are far greater than worrying about AI negatively. We still have a long way to go before it replaces creativity. On this day after Frank Gehry's passing, I note that he embraced technology as it actually helped him to be more creative. Let's do everything in our power to ensure it is used as a tool to make us better, and not let it be in charge. With every new capability we discover, some people will use it positively, while others will abuse it. It takes a far more creative person to use it to their advantage. 

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

Artificial Intelligence has actually been with us for some time in various early forms, including expert systems and even simple programming. Technology has been far more helpful than a hindrance, and I am sure each of us could cite many positive outcomes. In many forms, I hear that our creativity is at risk from the use of codes, standards, and even tools to support our ability to do our jobs. Our challenges are far greater than worrying about AI negatively. We still have a long way to go before it replaces creativity. On this day after Frank Gehry's passing, I note that he embraced technology as it actually helped him to be more creative. Let's do everything in our power to ensure it is used as a tool to make us better, and not let it be in charge. With every new capability we discover, some people will use it positively, while others will abuse it. It takes a far more creative person to use it to their advantage. 

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

Artificial Intelligence has actually been with us for some time in various early forms, including expert systems and even simple programming. Technology has been far more helpful than a hindrance, and I am sure each of us could cite many positive outcomes. In many forms, I hear that our creativity is at risk from the use of codes, standards, and even tools to support our ability to do our jobs. Our challenges are far greater than worrying about AI negatively. We still have a long way to go before it replaces creativity. On this day after Frank Gehry's passing, I note that he embraced technology as it actually helped him to be more creative. Let's do everything in our power to ensure it is used as a tool to make us better, and not let it be in charge. With every new capability we discover, some people will use it positively, while others will abuse it. It takes a far more creative person to use it to their advantage. 

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

Artificial Intelligence has actually been with us for some time in various early forms, including expert systems and even simple programming. Technology has been far more helpful than a hindrance, and I am sure each of us could cite many positive outcomes. In many forms, I hear that our creativity is at risk from the use of codes, standards, and even tools to support our ability to do our jobs. Our challenges are far greater than worrying about AI negatively. We still have a long way to go before it replaces creativity. On this day after Frank Gehry's passing, I note that he embraced technology as it actually helped him to be more creative. Let's do everything in our power to ensure it is used as a tool to make us better, and not let it be in charge. With every new capability we discover, some people will use it positively, while others will abuse it. It takes a far more creative person to use it to their advantage. 

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

Great power comes with great responsability 

:) 

------------------------------
Thesla Collier
HNTB

Original Message:
Sent: 12-09-2025 09:30 AM
From: Benjamin Horten, AIA
Subject: AI in the Studio: Friend or Foe?

Original Message:
Sent: 12/6/2025 1:32:00 PM
From: Dana Smith, FAIA Member Emeritus
Subject: RE: AI in the Studio: Friend or Foe?

Artificial Intelligence has actually been with us for some time in various early forms, including expert systems and even simple programming. Technology has been far more helpful than a hindrance, and I am sure each of us could cite many positive outcomes. In many forms, I hear that our creativity is at risk from the use of codes, standards, and even tools to support our ability to do our jobs. Our challenges are far greater than worrying about AI negatively. We still have a long way to go before it replaces creativity. On this day after Frank Gehry's passing, I note that he embraced technology as it actually helped him to be more creative. Let's do everything in our power to ensure it is used as a tool to make us better, and not let it be in charge. With every new capability we discover, some people will use it positively, while others will abuse it. It takes a far more creative person to use it to their advantage. 

------------------------------
Dana Smith FAIA Member Emeritus
Herndon VA

Original Message:
Sent: 12-04-2025 08:08 PM
From: Thesla Collier, Intl. Assoc. AIA
Subject: AI in the Studio: Friend or Foe?

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

------------------------------
Thesla Collier
HNTB
------------------------------

Artificial Intelligence has actually been with us for some time in various early forms, including expert systems and even simple programming. Technology has been far more helpful than a hindrance, and I am sure each of us could cite many positive outcomes. In many forms, I hear that our creativity is at risk from the use of codes, standards, and even tools to support our ability to do our jobs. Our challenges are far greater than worrying about AI negatively. We still have a long way to go before it replaces creativity. On this day after Frank Gehry's passing, I note that he embraced technology as it actually helped him to be more creative. Let's do everything in our power to ensure it is used as a tool to make us better, and not let it be in charge. With every new capability we discover, some people will use it positively, while others will abuse it. It takes a far more creative person to use it to their advantage. 

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

@Thesla Collier, Intl. Assoc. AIA I appreciate this conversation because it reflects where many architects are right now: curious, cautious, and trying to understand where AI fits into practice. But after presenting last week at NeurIPS in San Diego, the Conference on Neural Information Processing Systems, the world's largest and most influential AI research gathering, I can say this with confidence:

We are asking the wrong question.

While we debate whether AI is a "friend or foe" in the studio, NeurIPS brought together 22,000 researchers working on the systems that will shape the future of cities, mobility, infrastructure, and the built world.

Here is the critical insight:
NeurIPS is literally focused on building neural networks that describe the real world.
What they don't have, and what they openly said they need, is the neural network of buildings and cities themselves.

And that's where our profession is completely missing.

As an architect who has spent decades working on data, interoperability, and building intelligence, here's the hard truth:

The built environment is invisible to AI, not because the technology isn't ready, but because our profession refuses to expose the intelligence inside our buildings.

At Greenbuild this year, I heard the AIA Large Firm Roundtable (LFRT) say they strip data before handoff and avoid sharing models out of fear of losing IP. This is the same pattern we saw with CAD and BIM, and now the stakes are dramatically higher.

Meanwhile, at NeurIPS, I learned something the AEC industry has not yet absorbed:

Because architects don't share real operational intelligence, the AI industry is forced to use AI to create synthetic data about buildings and cities.

AI researchers are literally fabricating data for:

• building systems

• equipment behavior

• energy loads

• occupancy patterns

• maintenance events

• city infrastructure

• environmental interactions

Not because they want synthetic data, but because the real data is locked in PDFs, proprietary silos, and "means-and-methods" IP walls.

To AI researchers, our buildings appear:

• mute

• structureless

• disconnected

• uninterpretable

So AI does what it always does when data is missing:

It hallucinates the missing context and rebuilds our world from scratch.

This should concern us. It means the models shaping the future of cities do not contain architectural intelligence, because we have not provided it.

And here's the deeper issue:

If architects do not define machine-readable buildings and cities, AI will.

And it will do so without our expertise.

This is not about giving away creativity.
It's about giving away silence.

Shared, structured, machine-readable intelligence is one of the most sustainable materials we can contribute to the built world. If we continue treating data as proprietary instead of foundational, we hold back the very outcomes we claim to champion: climate performance, resilience, operational safety, great design, and long-term value for owners.

If you're interested in what I presented at NeurIPS, and what I saw there that directly affects our profession, I shared a summary here:

LinkedIn post:
https://www.linkedin.com/posts/kimononuma_neurips-urbanai-digitaltwins-activity-7404160879551799296-2eE-?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAErPBoBs61SLDRQoQjZHu1gU7qXr4VnM3I

AutomatedBuildings article:
https://www.automatedbuildings.com/2025/12/urban-ai-meets-the-built-environment-breaking-the-firewalls-that-keep-buildings-silent/

AI is enhancing creativity and replacing tasks, but it replaces only what we refuse to define.

Diagrams, specs, options, renderings, animations, 3D… AI can generate all of these synthetically because the real intelligence inside our buildings is locked behind our own walls.

Authorship remains where it always has:
in the architect's ability to give form, relationships, and meaning to the built environment, and now, to its data.

If we provide that foundation, AI becomes a collaborator. If we don't, AI will approximate the built world using synthetic models, and authorship simply shifts elsewhere.

AI is already reshaping our field.
The choice is whether we participate in shaping it or watch from the sidelines.

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

I see a lot of posts that boil down to "AI is inevitable, so just give in". It's got me wondering--why does it need to be inevitable? It's not an earthquake or a sunrise. It's something created by people. It only exists if we choose to make it exist. It only takes over our industry if we decide to let it. It sounds really defeatist to say that we have to accept the consequences of AI just because some billionaires are running around saying it has to happen.

Sam Altman (CEO of OpenAI) was on Jimmy Fallon last night, saying that he couldn't imagine raising his child without AI to refer to (to quote him: "I cannot imagine having gone through figuring out how to raise a newborn without ChatGPT"). It feels like he's decided to give up an essential piece of his humanity--figuring stuff out. Or, having an interaction with another person who can share their perspective and wisdom as part of a real, human relationship.

Giving all our data to these guys for free, just so that they can develop tools that we get to pay a monthly subscription price for doesn't seem like a fair trade. Copyright law exists for a reason. IP takes time and skill to develop. It's not like they're asking, either--they're out there scraping the internet. Why not stand up to them?

I 100% realize I sound like a cranky old man with this. It's ok, I own it. I'm all for new tools and ideas, but this one feels pretty un-thought-through to me. It feels like there's a lot of FOMO here that's driving hasty decisions and not much thinking ahead. If a client came to you with this kind of rushed enthusiasm, wouldn't you tell them to slow down and spend a little time planning?

------------------------------
Ian Toner AIA
Toner Architects
Philadelphia PA

Original Message:
Sent: 12-09-2025 11:14 AM
From: Kimon Onuma, FAIA
Subject: AI in the Studio: Friend or Foe?

@Thesla Collier, Intl. Assoc. AIA I appreciate this conversation because it reflects where many architects are right now: curious, cautious, and trying to understand where AI fits into practice. But after presenting last week at NeurIPS in San Diego, the Conference on Neural Information Processing Systems, the world's largest and most influential AI research gathering, I can say this with confidence:

We are asking the wrong question.

While we debate whether AI is a "friend or foe" in the studio, NeurIPS brought together 22,000 researchers working on the systems that will shape the future of cities, mobility, infrastructure, and the built world.

Here is the critical insight:
NeurIPS is literally focused on building neural networks that describe the real world.
What they don't have, and what they openly said they need, is the neural network of buildings and cities themselves.

And that's where our profession is completely missing.

As an architect who has spent decades working on data, interoperability, and building intelligence, here's the hard truth:

The built environment is invisible to AI, not because the technology isn't ready, but because our profession refuses to expose the intelligence inside our buildings.

At Greenbuild this year, I heard the AIA Large Firm Roundtable (LFRT) say they strip data before handoff and avoid sharing models out of fear of losing IP. This is the same pattern we saw with CAD and BIM, and now the stakes are dramatically higher.

Meanwhile, at NeurIPS, I learned something the AEC industry has not yet absorbed:

Because architects don't share real operational intelligence, the AI industry is forced to use AI to create synthetic data about buildings and cities.

AI researchers are literally fabricating data for:

• building systems

• equipment behavior

• energy loads

• occupancy patterns

• maintenance events

• city infrastructure

• environmental interactions

Not because they want synthetic data, but because the real data is locked in PDFs, proprietary silos, and "means-and-methods" IP walls.

To AI researchers, our buildings appear:

• mute

• structureless

• disconnected

• uninterpretable

So AI does what it always does when data is missing:

It hallucinates the missing context and rebuilds our world from scratch.

This should concern us. It means the models shaping the future of cities do not contain architectural intelligence, because we have not provided it.

And here's the deeper issue:

If architects do not define machine-readable buildings and cities, AI will.

And it will do so without our expertise.

This is not about giving away creativity.
It's about giving away silence.

Shared, structured, machine-readable intelligence is one of the most sustainable materials we can contribute to the built world. If we continue treating data as proprietary instead of foundational, we hold back the very outcomes we claim to champion: climate performance, resilience, operational safety, great design, and long-term value for owners.

If you're interested in what I presented at NeurIPS, and what I saw there that directly affects our profession, I shared a summary here:

LinkedIn post:
https://www.linkedin.com/posts/kimononuma_neurips-urbanai-digitaltwins-activity-7404160879551799296-2eE-?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAErPBoBs61SLDRQoQjZHu1gU7qXr4VnM3I

AutomatedBuildings article:
https://www.automatedbuildings.com/2025/12/urban-ai-meets-the-built-environment-breaking-the-firewalls-that-keep-buildings-silent/

AI is enhancing creativity and replacing tasks, but it replaces only what we refuse to define.

Diagrams, specs, options, renderings, animations, 3D… AI can generate all of these synthetically because the real intelligence inside our buildings is locked behind our own walls.

Authorship remains where it always has:
in the architect's ability to give form, relationships, and meaning to the built environment, and now, to its data.

If we provide that foundation, AI becomes a collaborator. If we don't, AI will approximate the built world using synthetic models, and authorship simply shifts elsewhere.

AI is already reshaping our field.
The choice is whether we participate in shaping it or watch from the sidelines.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-04-2025 08:08 PM
From: Thesla Collier, Intl. Assoc. AIA
Subject: AI in the Studio: Friend or Foe?

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

------------------------------
Thesla Collier
HNTB
------------------------------

Ian, I appreciate your perspective, and I share many of the same concerns. I've been concerned about the trajectory of AI for a long time, not because I think it's inherently good or bad, but because it's powerful. And when something powerful emerges, the worst response is to ignore it. The best response is to understand it deeply enough to guide it.

That's how I approached AI from the beginning. The more I learned, the more I realized that our industry can't shape something we refuse to engage with.

You brought up billionaires pushing AI forward. I don't think we should accept their vision as inevitable. In fact, I believe it's our obligation as architects to help guide how AI interacts with the built world. 

It reminds me of raising teenagers: We don't ignore them because we don't like what they're doing; we stay involved, try to understand their thinking, and guide them toward better outcomes. If we disengage, they simply go their own way.

AI is the same. It's already here, and it will continue to evolve. If we step back, others will shape how it understands buildings, cities, and human environments, and that may not align with our values at all.

And this connects to something more profound: the lack of useful, machine-readable information inside buildings didn't start with AI.

For decades, architects have overprotected data, drawings, models, specs, not out of malice, but out of fear: fear of losing IP, fear of being copied, fear of liability. That's understandable. But it means owners don't receive the operational intelligence their buildings need.

Now, AI is simply exposing that longstanding issue. When researchers try to apply AI to the built environment, they discover there's no consistent data, no semantics, no identities, no systems to read. So they create synthetic buildings instead. That isn't an AI problem; it's an architecture and industry problem that's finally coming into view.

And this brings me back to your original point: should AI be allowed to reshape our industry? My answer is: only if we refuse to participate.

If we stay engaged, understand the tools, and bring our knowledge of buildings, systems, and human environments into the conversation, we can help steer AI toward outcomes that support owners, the public, and the profession.

If we withdraw, AI won't stop. It will simply move forward without our input.

That's why I believe the question isn't "friend or foe." It's: How do we, as architects, show up to guide this in a direction that reflects our values?

I would really like to hear your thoughts, and others', on what that guidance should look like. How do we balance legitimate concerns about IP, craft, and authorship with the responsibility to help shape tools that are already impacting the world around us?

This is exactly the kind of conversation the AIA should be having.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA
------------------------------

Original Message:
Sent: 12-09-2025 05:47 PM
From: Ian Toner, AIA
Subject: AI in the Studio: Friend or Foe?

I see a lot of posts that boil down to "AI is inevitable, so just give in". It's got me wondering--why does it need to be inevitable? It's not an earthquake or a sunrise. It's something created by people. It only exists if we choose to make it exist. It only takes over our industry if we decide to let it. It sounds really defeatist to say that we have to accept the consequences of AI just because some billionaires are running around saying it has to happen.

Sam Altman (CEO of OpenAI) was on Jimmy Fallon last night, saying that he couldn't imagine raising his child without AI to refer to (to quote him: "I cannot imagine having gone through figuring out how to raise a newborn without ChatGPT"). It feels like he's decided to give up an essential piece of his humanity--figuring stuff out. Or, having an interaction with another person who can share their perspective and wisdom as part of a real, human relationship.

Giving all our data to these guys for free, just so that they can develop tools that we get to pay a monthly subscription price for doesn't seem like a fair trade. Copyright law exists for a reason. IP takes time and skill to develop. It's not like they're asking, either--they're out there scraping the internet. Why not stand up to them?

I 100% realize I sound like a cranky old man with this. It's ok, I own it. I'm all for new tools and ideas, but this one feels pretty un-thought-through to me. It feels like there's a lot of FOMO here that's driving hasty decisions and not much thinking ahead. If a client came to you with this kind of rushed enthusiasm, wouldn't you tell them to slow down and spend a little time planning?

------------------------------
Ian Toner AIA
Toner Architects
Philadelphia PA

Original Message:
Sent: 12-09-2025 11:14 AM
From: Kimon Onuma, FAIA
Subject: AI in the Studio: Friend or Foe?

@Thesla Collier, Intl. Assoc. AIA I appreciate this conversation because it reflects where many architects are right now: curious, cautious, and trying to understand where AI fits into practice. But after presenting last week at NeurIPS in San Diego, the Conference on Neural Information Processing Systems, the world's largest and most influential AI research gathering, I can say this with confidence:

We are asking the wrong question.

While we debate whether AI is a "friend or foe" in the studio, NeurIPS brought together 22,000 researchers working on the systems that will shape the future of cities, mobility, infrastructure, and the built world.

Here is the critical insight:
NeurIPS is literally focused on building neural networks that describe the real world.
What they don't have, and what they openly said they need, is the neural network of buildings and cities themselves.

And that's where our profession is completely missing.

As an architect who has spent decades working on data, interoperability, and building intelligence, here's the hard truth:

The built environment is invisible to AI, not because the technology isn't ready, but because our profession refuses to expose the intelligence inside our buildings.

At Greenbuild this year, I heard the AIA Large Firm Roundtable (LFRT) say they strip data before handoff and avoid sharing models out of fear of losing IP. This is the same pattern we saw with CAD and BIM, and now the stakes are dramatically higher.

Meanwhile, at NeurIPS, I learned something the AEC industry has not yet absorbed:

Because architects don't share real operational intelligence, the AI industry is forced to use AI to create synthetic data about buildings and cities.

AI researchers are literally fabricating data for:

• building systems

• equipment behavior

• energy loads

• occupancy patterns

• maintenance events

• city infrastructure

• environmental interactions

Not because they want synthetic data, but because the real data is locked in PDFs, proprietary silos, and "means-and-methods" IP walls.

To AI researchers, our buildings appear:

• mute

• structureless

• disconnected

• uninterpretable

So AI does what it always does when data is missing:

It hallucinates the missing context and rebuilds our world from scratch.

This should concern us. It means the models shaping the future of cities do not contain architectural intelligence, because we have not provided it.

And here's the deeper issue:

If architects do not define machine-readable buildings and cities, AI will.

And it will do so without our expertise.

This is not about giving away creativity.
It's about giving away silence.

Shared, structured, machine-readable intelligence is one of the most sustainable materials we can contribute to the built world. If we continue treating data as proprietary instead of foundational, we hold back the very outcomes we claim to champion: climate performance, resilience, operational safety, great design, and long-term value for owners.

If you're interested in what I presented at NeurIPS, and what I saw there that directly affects our profession, I shared a summary here:

LinkedIn post:
https://www.linkedin.com/posts/kimononuma_neurips-urbanai-digitaltwins-activity-7404160879551799296-2eE-?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAErPBoBs61SLDRQoQjZHu1gU7qXr4VnM3I

AutomatedBuildings article:
https://www.automatedbuildings.com/2025/12/urban-ai-meets-the-built-environment-breaking-the-firewalls-that-keep-buildings-silent/

AI is enhancing creativity and replacing tasks, but it replaces only what we refuse to define.

Diagrams, specs, options, renderings, animations, 3D… AI can generate all of these synthetically because the real intelligence inside our buildings is locked behind our own walls.

Authorship remains where it always has:
in the architect's ability to give form, relationships, and meaning to the built environment, and now, to its data.

If we provide that foundation, AI becomes a collaborator. If we don't, AI will approximate the built world using synthetic models, and authorship simply shifts elsewhere.

AI is already reshaping our field.
The choice is whether we participate in shaping it or watch from the sidelines.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-04-2025 08:08 PM
From: Thesla Collier, Intl. Assoc. AIA
Subject: AI in the Studio: Friend or Foe?

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

------------------------------
Thesla Collier
HNTB
------------------------------

Ian, I appreciate your perspective, and I share many of the same concerns. I've been concerned about the trajectory of AI for a long time, not because I think it's inherently good or bad, but because it's powerful. And when something powerful emerges, the worst response is to ignore it. The best response is to understand it deeply enough to guide it.

That's how I approached AI from the beginning. The more I learned, the more I realized that our industry can't shape something we refuse to engage with.

You brought up billionaires pushing AI forward. I don't think we should accept their vision as inevitable. In fact, I believe it's our obligation as architects to help guide how AI interacts with the built world. 

It reminds me of raising teenagers: We don't ignore them because we don't like what they're doing; we stay involved, try to understand their thinking, and guide them toward better outcomes. If we disengage, they simply go their own way.

AI is the same. It's already here, and it will continue to evolve. If we step back, others will shape how it understands buildings, cities, and human environments, and that may not align with our values at all.

And this connects to something more profound: the lack of useful, machine-readable information inside buildings didn't start with AI.

For decades, architects have overprotected data, drawings, models, specs, not out of malice, but out of fear: fear of losing IP, fear of being copied, fear of liability. That's understandable. But it means owners don't receive the operational intelligence their buildings need.

Now, AI is simply exposing that longstanding issue. When researchers try to apply AI to the built environment, they discover there's no consistent data, no semantics, no identities, no systems to read. So they create synthetic buildings instead. That isn't an AI problem; it's an architecture and industry problem that's finally coming into view.

And this brings me back to your original point: should AI be allowed to reshape our industry? My answer is: only if we refuse to participate.

If we stay engaged, understand the tools, and bring our knowledge of buildings, systems, and human environments into the conversation, we can help steer AI toward outcomes that support owners, the public, and the profession.

If we withdraw, AI won't stop. It will simply move forward without our input.

That's why I believe the question isn't "friend or foe." It's: How do we, as architects, show up to guide this in a direction that reflects our values?

I would really like to hear your thoughts, and others', on what that guidance should look like. How do we balance legitimate concerns about IP, craft, and authorship with the responsibility to help shape tools that are already impacting the world around us?

This is exactly the kind of conversation the AIA should be having.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-09-2025 05:47 PM
From: Ian Toner, AIA
Subject: AI in the Studio: Friend or Foe?

I see a lot of posts that boil down to "AI is inevitable, so just give in". It's got me wondering--why does it need to be inevitable? It's not an earthquake or a sunrise. It's something created by people. It only exists if we choose to make it exist. It only takes over our industry if we decide to let it. It sounds really defeatist to say that we have to accept the consequences of AI just because some billionaires are running around saying it has to happen.

Sam Altman (CEO of OpenAI) was on Jimmy Fallon last night, saying that he couldn't imagine raising his child without AI to refer to (to quote him: "I cannot imagine having gone through figuring out how to raise a newborn without ChatGPT"). It feels like he's decided to give up an essential piece of his humanity--figuring stuff out. Or, having an interaction with another person who can share their perspective and wisdom as part of a real, human relationship.

Giving all our data to these guys for free, just so that they can develop tools that we get to pay a monthly subscription price for doesn't seem like a fair trade. Copyright law exists for a reason. IP takes time and skill to develop. It's not like they're asking, either--they're out there scraping the internet. Why not stand up to them?

I 100% realize I sound like a cranky old man with this. It's ok, I own it. I'm all for new tools and ideas, but this one feels pretty un-thought-through to me. It feels like there's a lot of FOMO here that's driving hasty decisions and not much thinking ahead. If a client came to you with this kind of rushed enthusiasm, wouldn't you tell them to slow down and spend a little time planning?

------------------------------
Ian Toner AIA
Toner Architects
Philadelphia PA

Original Message:
Sent: 12-09-2025 11:14 AM
From: Kimon Onuma, FAIA
Subject: AI in the Studio: Friend or Foe?

@Thesla Collier, Intl. Assoc. AIA I appreciate this conversation because it reflects where many architects are right now: curious, cautious, and trying to understand where AI fits into practice. But after presenting last week at NeurIPS in San Diego, the Conference on Neural Information Processing Systems, the world's largest and most influential AI research gathering, I can say this with confidence:

We are asking the wrong question.

While we debate whether AI is a "friend or foe" in the studio, NeurIPS brought together 22,000 researchers working on the systems that will shape the future of cities, mobility, infrastructure, and the built world.

Here is the critical insight:
NeurIPS is literally focused on building neural networks that describe the real world.
What they don't have, and what they openly said they need, is the neural network of buildings and cities themselves.

And that's where our profession is completely missing.

As an architect who has spent decades working on data, interoperability, and building intelligence, here's the hard truth:

The built environment is invisible to AI, not because the technology isn't ready, but because our profession refuses to expose the intelligence inside our buildings.

At Greenbuild this year, I heard the AIA Large Firm Roundtable (LFRT) say they strip data before handoff and avoid sharing models out of fear of losing IP. This is the same pattern we saw with CAD and BIM, and now the stakes are dramatically higher.

Meanwhile, at NeurIPS, I learned something the AEC industry has not yet absorbed:

Because architects don't share real operational intelligence, the AI industry is forced to use AI to create synthetic data about buildings and cities.

AI researchers are literally fabricating data for:

• building systems

• equipment behavior

• energy loads

• occupancy patterns

• maintenance events

• city infrastructure

• environmental interactions

Not because they want synthetic data, but because the real data is locked in PDFs, proprietary silos, and "means-and-methods" IP walls.

To AI researchers, our buildings appear:

• mute

• structureless

• disconnected

• uninterpretable

So AI does what it always does when data is missing:

It hallucinates the missing context and rebuilds our world from scratch.

This should concern us. It means the models shaping the future of cities do not contain architectural intelligence, because we have not provided it.

And here's the deeper issue:

If architects do not define machine-readable buildings and cities, AI will.

And it will do so without our expertise.

This is not about giving away creativity.
It's about giving away silence.

Shared, structured, machine-readable intelligence is one of the most sustainable materials we can contribute to the built world. If we continue treating data as proprietary instead of foundational, we hold back the very outcomes we claim to champion: climate performance, resilience, operational safety, great design, and long-term value for owners.

If you're interested in what I presented at NeurIPS, and what I saw there that directly affects our profession, I shared a summary here:

LinkedIn post:
https://www.linkedin.com/posts/kimononuma_neurips-urbanai-digitaltwins-activity-7404160879551799296-2eE-?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAErPBoBs61SLDRQoQjZHu1gU7qXr4VnM3I

AutomatedBuildings article:
https://www.automatedbuildings.com/2025/12/urban-ai-meets-the-built-environment-breaking-the-firewalls-that-keep-buildings-silent/

AI is enhancing creativity and replacing tasks, but it replaces only what we refuse to define.

Diagrams, specs, options, renderings, animations, 3D… AI can generate all of these synthetically because the real intelligence inside our buildings is locked behind our own walls.

Authorship remains where it always has:
in the architect's ability to give form, relationships, and meaning to the built environment, and now, to its data.

If we provide that foundation, AI becomes a collaborator. If we don't, AI will approximate the built world using synthetic models, and authorship simply shifts elsewhere.

AI is already reshaping our field.
The choice is whether we participate in shaping it or watch from the sidelines.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-04-2025 08:08 PM
From: Thesla Collier, Intl. Assoc. AIA
Subject: AI in the Studio: Friend or Foe?

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

------------------------------
Thesla Collier
HNTB
------------------------------

Benjamin,  good points.

This actually goes back to a more basic issue. Architects are still delivering information as documents, not as usable data. Owners then store that material as analog records, just digitized PDFs and folders, and later wonder why it can't be used effectively.

This isn't an AI problem; it's a data and process problem. AI can help clean up some of the mess, but if we don't fix how we value, define, structure, and hand off building information in the first place, we're just automating bad practices. The first step is improving our process, AI or not.

Ian, I appreciate your perspective, and I share many of the same concerns. I've been concerned about the trajectory of AI for a long time, not because I think it's inherently good or bad, but because it's powerful. And when something powerful emerges, the worst response is to ignore it. The best response is to understand it deeply enough to guide it.

That's how I approached AI from the beginning. The more I learned, the more I realized that our industry can't shape something we refuse to engage with.

You brought up billionaires pushing AI forward. I don't think we should accept their vision as inevitable. In fact, I believe it's our obligation as architects to help guide how AI interacts with the built world. 

It reminds me of raising teenagers: We don't ignore them because we don't like what they're doing; we stay involved, try to understand their thinking, and guide them toward better outcomes. If we disengage, they simply go their own way.

AI is the same. It's already here, and it will continue to evolve. If we step back, others will shape how it understands buildings, cities, and human environments, and that may not align with our values at all.

And this connects to something more profound: the lack of useful, machine-readable information inside buildings didn't start with AI.

For decades, architects have overprotected data, drawings, models, specs, not out of malice, but out of fear: fear of losing IP, fear of being copied, fear of liability. That's understandable. But it means owners don't receive the operational intelligence their buildings need.

Now, AI is simply exposing that longstanding issue. When researchers try to apply AI to the built environment, they discover there's no consistent data, no semantics, no identities, no systems to read. So they create synthetic buildings instead. That isn't an AI problem; it's an architecture and industry problem that's finally coming into view.

And this brings me back to your original point: should AI be allowed to reshape our industry? My answer is: only if we refuse to participate.

If we stay engaged, understand the tools, and bring our knowledge of buildings, systems, and human environments into the conversation, we can help steer AI toward outcomes that support owners, the public, and the profession.

If we withdraw, AI won't stop. It will simply move forward without our input.

That's why I believe the question isn't "friend or foe." It's: How do we, as architects, show up to guide this in a direction that reflects our values?

I would really like to hear your thoughts, and others', on what that guidance should look like. How do we balance legitimate concerns about IP, craft, and authorship with the responsibility to help shape tools that are already impacting the world around us?

This is exactly the kind of conversation the AIA should be having.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-09-2025 05:47 PM
From: Ian Toner, AIA
Subject: AI in the Studio: Friend or Foe?

I see a lot of posts that boil down to "AI is inevitable, so just give in". It's got me wondering--why does it need to be inevitable? It's not an earthquake or a sunrise. It's something created by people. It only exists if we choose to make it exist. It only takes over our industry if we decide to let it. It sounds really defeatist to say that we have to accept the consequences of AI just because some billionaires are running around saying it has to happen.

Sam Altman (CEO of OpenAI) was on Jimmy Fallon last night, saying that he couldn't imagine raising his child without AI to refer to (to quote him: "I cannot imagine having gone through figuring out how to raise a newborn without ChatGPT"). It feels like he's decided to give up an essential piece of his humanity--figuring stuff out. Or, having an interaction with another person who can share their perspective and wisdom as part of a real, human relationship.

Giving all our data to these guys for free, just so that they can develop tools that we get to pay a monthly subscription price for doesn't seem like a fair trade. Copyright law exists for a reason. IP takes time and skill to develop. It's not like they're asking, either--they're out there scraping the internet. Why not stand up to them?

I 100% realize I sound like a cranky old man with this. It's ok, I own it. I'm all for new tools and ideas, but this one feels pretty un-thought-through to me. It feels like there's a lot of FOMO here that's driving hasty decisions and not much thinking ahead. If a client came to you with this kind of rushed enthusiasm, wouldn't you tell them to slow down and spend a little time planning?

------------------------------
Ian Toner AIA
Toner Architects
Philadelphia PA

Original Message:
Sent: 12-09-2025 11:14 AM
From: Kimon Onuma, FAIA
Subject: AI in the Studio: Friend or Foe?

@Thesla Collier, Intl. Assoc. AIA I appreciate this conversation because it reflects where many architects are right now: curious, cautious, and trying to understand where AI fits into practice. But after presenting last week at NeurIPS in San Diego, the Conference on Neural Information Processing Systems, the world's largest and most influential AI research gathering, I can say this with confidence:

We are asking the wrong question.

While we debate whether AI is a "friend or foe" in the studio, NeurIPS brought together 22,000 researchers working on the systems that will shape the future of cities, mobility, infrastructure, and the built world.

Here is the critical insight:
NeurIPS is literally focused on building neural networks that describe the real world.
What they don't have, and what they openly said they need, is the neural network of buildings and cities themselves.

And that's where our profession is completely missing.

As an architect who has spent decades working on data, interoperability, and building intelligence, here's the hard truth:

The built environment is invisible to AI, not because the technology isn't ready, but because our profession refuses to expose the intelligence inside our buildings.

At Greenbuild this year, I heard the AIA Large Firm Roundtable (LFRT) say they strip data before handoff and avoid sharing models out of fear of losing IP. This is the same pattern we saw with CAD and BIM, and now the stakes are dramatically higher.

Meanwhile, at NeurIPS, I learned something the AEC industry has not yet absorbed:

Because architects don't share real operational intelligence, the AI industry is forced to use AI to create synthetic data about buildings and cities.

AI researchers are literally fabricating data for:

• building systems

• equipment behavior

• energy loads

• occupancy patterns

• maintenance events

• city infrastructure

• environmental interactions

Not because they want synthetic data, but because the real data is locked in PDFs, proprietary silos, and "means-and-methods" IP walls.

To AI researchers, our buildings appear:

• mute

• structureless

• disconnected

• uninterpretable

So AI does what it always does when data is missing:

It hallucinates the missing context and rebuilds our world from scratch.

This should concern us. It means the models shaping the future of cities do not contain architectural intelligence, because we have not provided it.

And here's the deeper issue:

If architects do not define machine-readable buildings and cities, AI will.

And it will do so without our expertise.

This is not about giving away creativity.
It's about giving away silence.

Shared, structured, machine-readable intelligence is one of the most sustainable materials we can contribute to the built world. If we continue treating data as proprietary instead of foundational, we hold back the very outcomes we claim to champion: climate performance, resilience, operational safety, great design, and long-term value for owners.

If you're interested in what I presented at NeurIPS, and what I saw there that directly affects our profession, I shared a summary here:

LinkedIn post:
https://www.linkedin.com/posts/kimononuma_neurips-urbanai-digitaltwins-activity-7404160879551799296-2eE-?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAErPBoBs61SLDRQoQjZHu1gU7qXr4VnM3I

AutomatedBuildings article:
https://www.automatedbuildings.com/2025/12/urban-ai-meets-the-built-environment-breaking-the-firewalls-that-keep-buildings-silent/

AI is enhancing creativity and replacing tasks, but it replaces only what we refuse to define.

Diagrams, specs, options, renderings, animations, 3D… AI can generate all of these synthetically because the real intelligence inside our buildings is locked behind our own walls.

Authorship remains where it always has:
in the architect's ability to give form, relationships, and meaning to the built environment, and now, to its data.

If we provide that foundation, AI becomes a collaborator. If we don't, AI will approximate the built world using synthetic models, and authorship simply shifts elsewhere.

AI is already reshaping our field.
The choice is whether we participate in shaping it or watch from the sidelines.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-04-2025 08:08 PM
From: Thesla Collier, Intl. Assoc. AIA
Subject: AI in the Studio: Friend or Foe?

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

------------------------------
Thesla Collier
HNTB
------------------------------

Benjamin,  good points.

This actually goes back to a more basic issue. Architects are still delivering information as documents, not as usable data. Owners then store that material as analog records, just digitized PDFs and folders, and later wonder why it can't be used effectively.

This isn't an AI problem; it's a data and process problem. AI can help clean up some of the mess, but if we don't fix how we value, define, structure, and hand off building information in the first place, we're just automating bad practices. The first step is improving our process, AI or not.

Ian, I appreciate your perspective, and I share many of the same concerns. I've been concerned about the trajectory of AI for a long time, not because I think it's inherently good or bad, but because it's powerful. And when something powerful emerges, the worst response is to ignore it. The best response is to understand it deeply enough to guide it.

That's how I approached AI from the beginning. The more I learned, the more I realized that our industry can't shape something we refuse to engage with.

You brought up billionaires pushing AI forward. I don't think we should accept their vision as inevitable. In fact, I believe it's our obligation as architects to help guide how AI interacts with the built world. 

It reminds me of raising teenagers: We don't ignore them because we don't like what they're doing; we stay involved, try to understand their thinking, and guide them toward better outcomes. If we disengage, they simply go their own way.

AI is the same. It's already here, and it will continue to evolve. If we step back, others will shape how it understands buildings, cities, and human environments, and that may not align with our values at all.

And this connects to something more profound: the lack of useful, machine-readable information inside buildings didn't start with AI.

For decades, architects have overprotected data, drawings, models, specs, not out of malice, but out of fear: fear of losing IP, fear of being copied, fear of liability. That's understandable. But it means owners don't receive the operational intelligence their buildings need.

Now, AI is simply exposing that longstanding issue. When researchers try to apply AI to the built environment, they discover there's no consistent data, no semantics, no identities, no systems to read. So they create synthetic buildings instead. That isn't an AI problem; it's an architecture and industry problem that's finally coming into view.

And this brings me back to your original point: should AI be allowed to reshape our industry? My answer is: only if we refuse to participate.

If we stay engaged, understand the tools, and bring our knowledge of buildings, systems, and human environments into the conversation, we can help steer AI toward outcomes that support owners, the public, and the profession.

If we withdraw, AI won't stop. It will simply move forward without our input.

That's why I believe the question isn't "friend or foe." It's: How do we, as architects, show up to guide this in a direction that reflects our values?

I would really like to hear your thoughts, and others', on what that guidance should look like. How do we balance legitimate concerns about IP, craft, and authorship with the responsibility to help shape tools that are already impacting the world around us?

This is exactly the kind of conversation the AIA should be having.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-09-2025 05:47 PM
From: Ian Toner, AIA
Subject: AI in the Studio: Friend or Foe?

I see a lot of posts that boil down to "AI is inevitable, so just give in". It's got me wondering--why does it need to be inevitable? It's not an earthquake or a sunrise. It's something created by people. It only exists if we choose to make it exist. It only takes over our industry if we decide to let it. It sounds really defeatist to say that we have to accept the consequences of AI just because some billionaires are running around saying it has to happen.

Sam Altman (CEO of OpenAI) was on Jimmy Fallon last night, saying that he couldn't imagine raising his child without AI to refer to (to quote him: "I cannot imagine having gone through figuring out how to raise a newborn without ChatGPT"). It feels like he's decided to give up an essential piece of his humanity--figuring stuff out. Or, having an interaction with another person who can share their perspective and wisdom as part of a real, human relationship.

Giving all our data to these guys for free, just so that they can develop tools that we get to pay a monthly subscription price for doesn't seem like a fair trade. Copyright law exists for a reason. IP takes time and skill to develop. It's not like they're asking, either--they're out there scraping the internet. Why not stand up to them?

I 100% realize I sound like a cranky old man with this. It's ok, I own it. I'm all for new tools and ideas, but this one feels pretty un-thought-through to me. It feels like there's a lot of FOMO here that's driving hasty decisions and not much thinking ahead. If a client came to you with this kind of rushed enthusiasm, wouldn't you tell them to slow down and spend a little time planning?

------------------------------
Ian Toner AIA
Toner Architects
Philadelphia PA

Original Message:
Sent: 12-09-2025 11:14 AM
From: Kimon Onuma, FAIA
Subject: AI in the Studio: Friend or Foe?

@Thesla Collier, Intl. Assoc. AIA I appreciate this conversation because it reflects where many architects are right now: curious, cautious, and trying to understand where AI fits into practice. But after presenting last week at NeurIPS in San Diego, the Conference on Neural Information Processing Systems, the world's largest and most influential AI research gathering, I can say this with confidence:

We are asking the wrong question.

While we debate whether AI is a "friend or foe" in the studio, NeurIPS brought together 22,000 researchers working on the systems that will shape the future of cities, mobility, infrastructure, and the built world.

Here is the critical insight:
NeurIPS is literally focused on building neural networks that describe the real world.
What they don't have, and what they openly said they need, is the neural network of buildings and cities themselves.

And that's where our profession is completely missing.

As an architect who has spent decades working on data, interoperability, and building intelligence, here's the hard truth:

The built environment is invisible to AI, not because the technology isn't ready, but because our profession refuses to expose the intelligence inside our buildings.

At Greenbuild this year, I heard the AIA Large Firm Roundtable (LFRT) say they strip data before handoff and avoid sharing models out of fear of losing IP. This is the same pattern we saw with CAD and BIM, and now the stakes are dramatically higher.

Meanwhile, at NeurIPS, I learned something the AEC industry has not yet absorbed:

Because architects don't share real operational intelligence, the AI industry is forced to use AI to create synthetic data about buildings and cities.

AI researchers are literally fabricating data for:

• building systems

• equipment behavior

• energy loads

• occupancy patterns

• maintenance events

• city infrastructure

• environmental interactions

Not because they want synthetic data, but because the real data is locked in PDFs, proprietary silos, and "means-and-methods" IP walls.

To AI researchers, our buildings appear:

• mute

• structureless

• disconnected

• uninterpretable

So AI does what it always does when data is missing:

It hallucinates the missing context and rebuilds our world from scratch.

This should concern us. It means the models shaping the future of cities do not contain architectural intelligence, because we have not provided it.

And here's the deeper issue:

If architects do not define machine-readable buildings and cities, AI will.

And it will do so without our expertise.

This is not about giving away creativity.
It's about giving away silence.

Shared, structured, machine-readable intelligence is one of the most sustainable materials we can contribute to the built world. If we continue treating data as proprietary instead of foundational, we hold back the very outcomes we claim to champion: climate performance, resilience, operational safety, great design, and long-term value for owners.

If you're interested in what I presented at NeurIPS, and what I saw there that directly affects our profession, I shared a summary here:

LinkedIn post:
https://www.linkedin.com/posts/kimononuma_neurips-urbanai-digitaltwins-activity-7404160879551799296-2eE-?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAErPBoBs61SLDRQoQjZHu1gU7qXr4VnM3I

AutomatedBuildings article:
https://www.automatedbuildings.com/2025/12/urban-ai-meets-the-built-environment-breaking-the-firewalls-that-keep-buildings-silent/

AI is enhancing creativity and replacing tasks, but it replaces only what we refuse to define.

Diagrams, specs, options, renderings, animations, 3D… AI can generate all of these synthetically because the real intelligence inside our buildings is locked behind our own walls.

Authorship remains where it always has:
in the architect's ability to give form, relationships, and meaning to the built environment, and now, to its data.

If we provide that foundation, AI becomes a collaborator. If we don't, AI will approximate the built world using synthetic models, and authorship simply shifts elsewhere.

AI is already reshaping our field.
The choice is whether we participate in shaping it or watch from the sidelines.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-04-2025 08:08 PM
From: Thesla Collier, Intl. Assoc. AIA
Subject: AI in the Studio: Friend or Foe?

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

------------------------------
Thesla Collier
HNTB
------------------------------

Ian, I appreciate your perspective, and I share many of the same concerns. I've been concerned about the trajectory of AI for a long time, not because I think it's inherently good or bad, but because it's powerful. And when something powerful emerges, the worst response is to ignore it. The best response is to understand it deeply enough to guide it.

That's how I approached AI from the beginning. The more I learned, the more I realized that our industry can't shape something we refuse to engage with.

You brought up billionaires pushing AI forward. I don't think we should accept their vision as inevitable. In fact, I believe it's our obligation as architects to help guide how AI interacts with the built world. 

It reminds me of raising teenagers: We don't ignore them because we don't like what they're doing; we stay involved, try to understand their thinking, and guide them toward better outcomes. If we disengage, they simply go their own way.

AI is the same. It's already here, and it will continue to evolve. If we step back, others will shape how it understands buildings, cities, and human environments, and that may not align with our values at all.

And this connects to something more profound: the lack of useful, machine-readable information inside buildings didn't start with AI.

For decades, architects have overprotected data, drawings, models, specs, not out of malice, but out of fear: fear of losing IP, fear of being copied, fear of liability. That's understandable. But it means owners don't receive the operational intelligence their buildings need.

Now, AI is simply exposing that longstanding issue. When researchers try to apply AI to the built environment, they discover there's no consistent data, no semantics, no identities, no systems to read. So they create synthetic buildings instead. That isn't an AI problem; it's an architecture and industry problem that's finally coming into view.

And this brings me back to your original point: should AI be allowed to reshape our industry? My answer is: only if we refuse to participate.

If we stay engaged, understand the tools, and bring our knowledge of buildings, systems, and human environments into the conversation, we can help steer AI toward outcomes that support owners, the public, and the profession.

If we withdraw, AI won't stop. It will simply move forward without our input.

That's why I believe the question isn't "friend or foe." It's: How do we, as architects, show up to guide this in a direction that reflects our values?

I would really like to hear your thoughts, and others', on what that guidance should look like. How do we balance legitimate concerns about IP, craft, and authorship with the responsibility to help shape tools that are already impacting the world around us?

This is exactly the kind of conversation the AIA should be having.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-09-2025 05:47 PM
From: Ian Toner, AIA
Subject: AI in the Studio: Friend or Foe?

I see a lot of posts that boil down to "AI is inevitable, so just give in". It's got me wondering--why does it need to be inevitable? It's not an earthquake or a sunrise. It's something created by people. It only exists if we choose to make it exist. It only takes over our industry if we decide to let it. It sounds really defeatist to say that we have to accept the consequences of AI just because some billionaires are running around saying it has to happen.

Sam Altman (CEO of OpenAI) was on Jimmy Fallon last night, saying that he couldn't imagine raising his child without AI to refer to (to quote him: "I cannot imagine having gone through figuring out how to raise a newborn without ChatGPT"). It feels like he's decided to give up an essential piece of his humanity--figuring stuff out. Or, having an interaction with another person who can share their perspective and wisdom as part of a real, human relationship.

Giving all our data to these guys for free, just so that they can develop tools that we get to pay a monthly subscription price for doesn't seem like a fair trade. Copyright law exists for a reason. IP takes time and skill to develop. It's not like they're asking, either--they're out there scraping the internet. Why not stand up to them?

I 100% realize I sound like a cranky old man with this. It's ok, I own it. I'm all for new tools and ideas, but this one feels pretty un-thought-through to me. It feels like there's a lot of FOMO here that's driving hasty decisions and not much thinking ahead. If a client came to you with this kind of rushed enthusiasm, wouldn't you tell them to slow down and spend a little time planning?

------------------------------
Ian Toner AIA
Toner Architects
Philadelphia PA

Original Message:
Sent: 12-09-2025 11:14 AM
From: Kimon Onuma, FAIA
Subject: AI in the Studio: Friend or Foe?

@Thesla Collier, Intl. Assoc. AIA I appreciate this conversation because it reflects where many architects are right now: curious, cautious, and trying to understand where AI fits into practice. But after presenting last week at NeurIPS in San Diego, the Conference on Neural Information Processing Systems, the world's largest and most influential AI research gathering, I can say this with confidence:

We are asking the wrong question.

While we debate whether AI is a "friend or foe" in the studio, NeurIPS brought together 22,000 researchers working on the systems that will shape the future of cities, mobility, infrastructure, and the built world.

Here is the critical insight:
NeurIPS is literally focused on building neural networks that describe the real world.
What they don't have, and what they openly said they need, is the neural network of buildings and cities themselves.

And that's where our profession is completely missing.

As an architect who has spent decades working on data, interoperability, and building intelligence, here's the hard truth:

The built environment is invisible to AI, not because the technology isn't ready, but because our profession refuses to expose the intelligence inside our buildings.

At Greenbuild this year, I heard the AIA Large Firm Roundtable (LFRT) say they strip data before handoff and avoid sharing models out of fear of losing IP. This is the same pattern we saw with CAD and BIM, and now the stakes are dramatically higher.

Meanwhile, at NeurIPS, I learned something the AEC industry has not yet absorbed:

Because architects don't share real operational intelligence, the AI industry is forced to use AI to create synthetic data about buildings and cities.

AI researchers are literally fabricating data for:

• building systems

• equipment behavior

• energy loads

• occupancy patterns

• maintenance events

• city infrastructure

• environmental interactions

Not because they want synthetic data, but because the real data is locked in PDFs, proprietary silos, and "means-and-methods" IP walls.

To AI researchers, our buildings appear:

• mute

• structureless

• disconnected

• uninterpretable

So AI does what it always does when data is missing:

It hallucinates the missing context and rebuilds our world from scratch.

This should concern us. It means the models shaping the future of cities do not contain architectural intelligence, because we have not provided it.

And here's the deeper issue:

If architects do not define machine-readable buildings and cities, AI will.

And it will do so without our expertise.

This is not about giving away creativity.
It's about giving away silence.

Shared, structured, machine-readable intelligence is one of the most sustainable materials we can contribute to the built world. If we continue treating data as proprietary instead of foundational, we hold back the very outcomes we claim to champion: climate performance, resilience, operational safety, great design, and long-term value for owners.

If you're interested in what I presented at NeurIPS, and what I saw there that directly affects our profession, I shared a summary here:

LinkedIn post:
https://www.linkedin.com/posts/kimononuma_neurips-urbanai-digitaltwins-activity-7404160879551799296-2eE-?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAErPBoBs61SLDRQoQjZHu1gU7qXr4VnM3I

AutomatedBuildings article:
https://www.automatedbuildings.com/2025/12/urban-ai-meets-the-built-environment-breaking-the-firewalls-that-keep-buildings-silent/

AI is enhancing creativity and replacing tasks, but it replaces only what we refuse to define.

Diagrams, specs, options, renderings, animations, 3D… AI can generate all of these synthetically because the real intelligence inside our buildings is locked behind our own walls.

Authorship remains where it always has:
in the architect's ability to give form, relationships, and meaning to the built environment, and now, to its data.

If we provide that foundation, AI becomes a collaborator. If we don't, AI will approximate the built world using synthetic models, and authorship simply shifts elsewhere.

AI is already reshaping our field.
The choice is whether we participate in shaping it or watch from the sidelines.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-04-2025 08:08 PM
From: Thesla Collier, Intl. Assoc. AIA
Subject: AI in the Studio: Friend or Foe?

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

------------------------------
Thesla Collier
HNTB
------------------------------

 Kimon:

Ian, I appreciate your perspective, and I share many of the same concerns. I've been concerned about the trajectory of AI for a long time, not because I think it's inherently good or bad, but because it's powerful. And when something powerful emerges, the worst response is to ignore it. The best response is to understand it deeply enough to guide it.

That's how I approached AI from the beginning. The more I learned, the more I realized that our industry can't shape something we refuse to engage with.

You brought up billionaires pushing AI forward. I don't think we should accept their vision as inevitable. In fact, I believe it's our obligation as architects to help guide how AI interacts with the built world. 

It reminds me of raising teenagers: We don't ignore them because we don't like what they're doing; we stay involved, try to understand their thinking, and guide them toward better outcomes. If we disengage, they simply go their own way.

AI is the same. It's already here, and it will continue to evolve. If we step back, others will shape how it understands buildings, cities, and human environments, and that may not align with our values at all.

And this connects to something more profound: the lack of useful, machine-readable information inside buildings didn't start with AI.

For decades, architects have overprotected data, drawings, models, specs, not out of malice, but out of fear: fear of losing IP, fear of being copied, fear of liability. That's understandable. But it means owners don't receive the operational intelligence their buildings need.

Now, AI is simply exposing that longstanding issue. When researchers try to apply AI to the built environment, they discover there's no consistent data, no semantics, no identities, no systems to read. So they create synthetic buildings instead. That isn't an AI problem; it's an architecture and industry problem that's finally coming into view.

And this brings me back to your original point: should AI be allowed to reshape our industry? My answer is: only if we refuse to participate.

If we stay engaged, understand the tools, and bring our knowledge of buildings, systems, and human environments into the conversation, we can help steer AI toward outcomes that support owners, the public, and the profession.

If we withdraw, AI won't stop. It will simply move forward without our input.

That's why I believe the question isn't "friend or foe." It's: How do we, as architects, show up to guide this in a direction that reflects our values?

I would really like to hear your thoughts, and others', on what that guidance should look like. How do we balance legitimate concerns about IP, craft, and authorship with the responsibility to help shape tools that are already impacting the world around us?

This is exactly the kind of conversation the AIA should be having.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-09-2025 05:47 PM
From: Ian Toner, AIA
Subject: AI in the Studio: Friend or Foe?

I see a lot of posts that boil down to "AI is inevitable, so just give in". It's got me wondering--why does it need to be inevitable? It's not an earthquake or a sunrise. It's something created by people. It only exists if we choose to make it exist. It only takes over our industry if we decide to let it. It sounds really defeatist to say that we have to accept the consequences of AI just because some billionaires are running around saying it has to happen.

Sam Altman (CEO of OpenAI) was on Jimmy Fallon last night, saying that he couldn't imagine raising his child without AI to refer to (to quote him: "I cannot imagine having gone through figuring out how to raise a newborn without ChatGPT"). It feels like he's decided to give up an essential piece of his humanity--figuring stuff out. Or, having an interaction with another person who can share their perspective and wisdom as part of a real, human relationship.

Giving all our data to these guys for free, just so that they can develop tools that we get to pay a monthly subscription price for doesn't seem like a fair trade. Copyright law exists for a reason. IP takes time and skill to develop. It's not like they're asking, either--they're out there scraping the internet. Why not stand up to them?

I 100% realize I sound like a cranky old man with this. It's ok, I own it. I'm all for new tools and ideas, but this one feels pretty un-thought-through to me. It feels like there's a lot of FOMO here that's driving hasty decisions and not much thinking ahead. If a client came to you with this kind of rushed enthusiasm, wouldn't you tell them to slow down and spend a little time planning?

------------------------------
Ian Toner AIA
Toner Architects
Philadelphia PA

Original Message:
Sent: 12-09-2025 11:14 AM
From: Kimon Onuma, FAIA
Subject: AI in the Studio: Friend or Foe?

@Thesla Collier, Intl. Assoc. AIA I appreciate this conversation because it reflects where many architects are right now: curious, cautious, and trying to understand where AI fits into practice. But after presenting last week at NeurIPS in San Diego, the Conference on Neural Information Processing Systems, the world's largest and most influential AI research gathering, I can say this with confidence:

We are asking the wrong question.

While we debate whether AI is a "friend or foe" in the studio, NeurIPS brought together 22,000 researchers working on the systems that will shape the future of cities, mobility, infrastructure, and the built world.

Here is the critical insight:
NeurIPS is literally focused on building neural networks that describe the real world.
What they don't have, and what they openly said they need, is the neural network of buildings and cities themselves.

And that's where our profession is completely missing.

As an architect who has spent decades working on data, interoperability, and building intelligence, here's the hard truth:

The built environment is invisible to AI, not because the technology isn't ready, but because our profession refuses to expose the intelligence inside our buildings.

At Greenbuild this year, I heard the AIA Large Firm Roundtable (LFRT) say they strip data before handoff and avoid sharing models out of fear of losing IP. This is the same pattern we saw with CAD and BIM, and now the stakes are dramatically higher.

Meanwhile, at NeurIPS, I learned something the AEC industry has not yet absorbed:

Because architects don't share real operational intelligence, the AI industry is forced to use AI to create synthetic data about buildings and cities.

AI researchers are literally fabricating data for:

• building systems

• equipment behavior

• energy loads

• occupancy patterns

• maintenance events

• city infrastructure

• environmental interactions

Not because they want synthetic data, but because the real data is locked in PDFs, proprietary silos, and "means-and-methods" IP walls.

To AI researchers, our buildings appear:

• mute

• structureless

• disconnected

• uninterpretable

So AI does what it always does when data is missing:

It hallucinates the missing context and rebuilds our world from scratch.

This should concern us. It means the models shaping the future of cities do not contain architectural intelligence, because we have not provided it.

And here's the deeper issue:

If architects do not define machine-readable buildings and cities, AI will.

And it will do so without our expertise.

This is not about giving away creativity.
It's about giving away silence.

Shared, structured, machine-readable intelligence is one of the most sustainable materials we can contribute to the built world. If we continue treating data as proprietary instead of foundational, we hold back the very outcomes we claim to champion: climate performance, resilience, operational safety, great design, and long-term value for owners.

If you're interested in what I presented at NeurIPS, and what I saw there that directly affects our profession, I shared a summary here:

LinkedIn post:
https://www.linkedin.com/posts/kimononuma_neurips-urbanai-digitaltwins-activity-7404160879551799296-2eE-?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAErPBoBs61SLDRQoQjZHu1gU7qXr4VnM3I

AutomatedBuildings article:
https://www.automatedbuildings.com/2025/12/urban-ai-meets-the-built-environment-breaking-the-firewalls-that-keep-buildings-silent/

AI is enhancing creativity and replacing tasks, but it replaces only what we refuse to define.

Diagrams, specs, options, renderings, animations, 3D… AI can generate all of these synthetically because the real intelligence inside our buildings is locked behind our own walls.

Authorship remains where it always has:
in the architect's ability to give form, relationships, and meaning to the built environment, and now, to its data.

If we provide that foundation, AI becomes a collaborator. If we don't, AI will approximate the built world using synthetic models, and authorship simply shifts elsewhere.

AI is already reshaping our field.
The choice is whether we participate in shaping it or watch from the sidelines.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-04-2025 08:08 PM
From: Thesla Collier, Intl. Assoc. AIA
Subject: AI in the Studio: Friend or Foe?

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

------------------------------
Thesla Collier
HNTB
------------------------------

I agree, trust is fundamental and humans must be in the loop. But I think trust comes less from the tool itself and more from clarity around what we're feeding it and what we expect in return.

If our workflows are still document-centric and ambiguous, it's hard to trust any system, AI or otherwise, because the inputs are opaque and the outcomes unpredictable. When building information is well defined, structured, and intentional, trust follows naturally.

In that sense, AI doesn't create trust on its own, it reflects the quality and discipline of the processes we already have.

 Kimon:

Ian, I appreciate your perspective, and I share many of the same concerns. I've been concerned about the trajectory of AI for a long time, not because I think it's inherently good or bad, but because it's powerful. And when something powerful emerges, the worst response is to ignore it. The best response is to understand it deeply enough to guide it.

That's how I approached AI from the beginning. The more I learned, the more I realized that our industry can't shape something we refuse to engage with.

You brought up billionaires pushing AI forward. I don't think we should accept their vision as inevitable. In fact, I believe it's our obligation as architects to help guide how AI interacts with the built world. 

It reminds me of raising teenagers: We don't ignore them because we don't like what they're doing; we stay involved, try to understand their thinking, and guide them toward better outcomes. If we disengage, they simply go their own way.

AI is the same. It's already here, and it will continue to evolve. If we step back, others will shape how it understands buildings, cities, and human environments, and that may not align with our values at all.

And this connects to something more profound: the lack of useful, machine-readable information inside buildings didn't start with AI.

For decades, architects have overprotected data, drawings, models, specs, not out of malice, but out of fear: fear of losing IP, fear of being copied, fear of liability. That's understandable. But it means owners don't receive the operational intelligence their buildings need.

Now, AI is simply exposing that longstanding issue. When researchers try to apply AI to the built environment, they discover there's no consistent data, no semantics, no identities, no systems to read. So they create synthetic buildings instead. That isn't an AI problem; it's an architecture and industry problem that's finally coming into view.

And this brings me back to your original point: should AI be allowed to reshape our industry? My answer is: only if we refuse to participate.

If we stay engaged, understand the tools, and bring our knowledge of buildings, systems, and human environments into the conversation, we can help steer AI toward outcomes that support owners, the public, and the profession.

If we withdraw, AI won't stop. It will simply move forward without our input.

That's why I believe the question isn't "friend or foe." It's: How do we, as architects, show up to guide this in a direction that reflects our values?

I would really like to hear your thoughts, and others', on what that guidance should look like. How do we balance legitimate concerns about IP, craft, and authorship with the responsibility to help shape tools that are already impacting the world around us?

This is exactly the kind of conversation the AIA should be having.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-09-2025 05:47 PM
From: Ian Toner, AIA
Subject: AI in the Studio: Friend or Foe?

I see a lot of posts that boil down to "AI is inevitable, so just give in". It's got me wondering--why does it need to be inevitable? It's not an earthquake or a sunrise. It's something created by people. It only exists if we choose to make it exist. It only takes over our industry if we decide to let it. It sounds really defeatist to say that we have to accept the consequences of AI just because some billionaires are running around saying it has to happen.

Sam Altman (CEO of OpenAI) was on Jimmy Fallon last night, saying that he couldn't imagine raising his child without AI to refer to (to quote him: "I cannot imagine having gone through figuring out how to raise a newborn without ChatGPT"). It feels like he's decided to give up an essential piece of his humanity--figuring stuff out. Or, having an interaction with another person who can share their perspective and wisdom as part of a real, human relationship.

Giving all our data to these guys for free, just so that they can develop tools that we get to pay a monthly subscription price for doesn't seem like a fair trade. Copyright law exists for a reason. IP takes time and skill to develop. It's not like they're asking, either--they're out there scraping the internet. Why not stand up to them?

I 100% realize I sound like a cranky old man with this. It's ok, I own it. I'm all for new tools and ideas, but this one feels pretty un-thought-through to me. It feels like there's a lot of FOMO here that's driving hasty decisions and not much thinking ahead. If a client came to you with this kind of rushed enthusiasm, wouldn't you tell them to slow down and spend a little time planning?

------------------------------
Ian Toner AIA
Toner Architects
Philadelphia PA

Original Message:
Sent: 12-09-2025 11:14 AM
From: Kimon Onuma, FAIA
Subject: AI in the Studio: Friend or Foe?

@Thesla Collier, Intl. Assoc. AIA I appreciate this conversation because it reflects where many architects are right now: curious, cautious, and trying to understand where AI fits into practice. But after presenting last week at NeurIPS in San Diego, the Conference on Neural Information Processing Systems, the world's largest and most influential AI research gathering, I can say this with confidence:

We are asking the wrong question.

While we debate whether AI is a "friend or foe" in the studio, NeurIPS brought together 22,000 researchers working on the systems that will shape the future of cities, mobility, infrastructure, and the built world.

Here is the critical insight:
NeurIPS is literally focused on building neural networks that describe the real world.
What they don't have, and what they openly said they need, is the neural network of buildings and cities themselves.

And that's where our profession is completely missing.

As an architect who has spent decades working on data, interoperability, and building intelligence, here's the hard truth:

The built environment is invisible to AI, not because the technology isn't ready, but because our profession refuses to expose the intelligence inside our buildings.

At Greenbuild this year, I heard the AIA Large Firm Roundtable (LFRT) say they strip data before handoff and avoid sharing models out of fear of losing IP. This is the same pattern we saw with CAD and BIM, and now the stakes are dramatically higher.

Meanwhile, at NeurIPS, I learned something the AEC industry has not yet absorbed:

Because architects don't share real operational intelligence, the AI industry is forced to use AI to create synthetic data about buildings and cities.

AI researchers are literally fabricating data for:

• building systems

• equipment behavior

• energy loads

• occupancy patterns

• maintenance events

• city infrastructure

• environmental interactions

Not because they want synthetic data, but because the real data is locked in PDFs, proprietary silos, and "means-and-methods" IP walls.

To AI researchers, our buildings appear:

• mute

• structureless

• disconnected

• uninterpretable

So AI does what it always does when data is missing:

It hallucinates the missing context and rebuilds our world from scratch.

This should concern us. It means the models shaping the future of cities do not contain architectural intelligence, because we have not provided it.

And here's the deeper issue:

If architects do not define machine-readable buildings and cities, AI will.

And it will do so without our expertise.

This is not about giving away creativity.
It's about giving away silence.

Shared, structured, machine-readable intelligence is one of the most sustainable materials we can contribute to the built world. If we continue treating data as proprietary instead of foundational, we hold back the very outcomes we claim to champion: climate performance, resilience, operational safety, great design, and long-term value for owners.

If you're interested in what I presented at NeurIPS, and what I saw there that directly affects our profession, I shared a summary here:

LinkedIn post:
https://www.linkedin.com/posts/kimononuma_neurips-urbanai-digitaltwins-activity-7404160879551799296-2eE-?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAErPBoBs61SLDRQoQjZHu1gU7qXr4VnM3I

AutomatedBuildings article:
https://www.automatedbuildings.com/2025/12/urban-ai-meets-the-built-environment-breaking-the-firewalls-that-keep-buildings-silent/

AI is enhancing creativity and replacing tasks, but it replaces only what we refuse to define.

Diagrams, specs, options, renderings, animations, 3D… AI can generate all of these synthetically because the real intelligence inside our buildings is locked behind our own walls.

Authorship remains where it always has:
in the architect's ability to give form, relationships, and meaning to the built environment, and now, to its data.

If we provide that foundation, AI becomes a collaborator. If we don't, AI will approximate the built world using synthetic models, and authorship simply shifts elsewhere.

AI is already reshaping our field.
The choice is whether we participate in shaping it or watch from the sidelines.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-04-2025 08:08 PM
From: Thesla Collier, Intl. Assoc. AIA
Subject: AI in the Studio: Friend or Foe?

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

------------------------------
Thesla Collier
HNTB
------------------------------

100%

I agree, trust is fundamental and humans must be in the loop. But I think trust comes less from the tool itself and more from clarity around what we're feeding it and what we expect in return.

If our workflows are still document-centric and ambiguous, it's hard to trust any system, AI or otherwise, because the inputs are opaque and the outcomes unpredictable. When building information is well defined, structured, and intentional, trust follows naturally.

In that sense, AI doesn't create trust on its own, it reflects the quality and discipline of the processes we already have.

 Kimon:

Ian, I appreciate your perspective, and I share many of the same concerns. I've been concerned about the trajectory of AI for a long time, not because I think it's inherently good or bad, but because it's powerful. And when something powerful emerges, the worst response is to ignore it. The best response is to understand it deeply enough to guide it.

That's how I approached AI from the beginning. The more I learned, the more I realized that our industry can't shape something we refuse to engage with.

You brought up billionaires pushing AI forward. I don't think we should accept their vision as inevitable. In fact, I believe it's our obligation as architects to help guide how AI interacts with the built world. 

It reminds me of raising teenagers: We don't ignore them because we don't like what they're doing; we stay involved, try to understand their thinking, and guide them toward better outcomes. If we disengage, they simply go their own way.

AI is the same. It's already here, and it will continue to evolve. If we step back, others will shape how it understands buildings, cities, and human environments, and that may not align with our values at all.

And this connects to something more profound: the lack of useful, machine-readable information inside buildings didn't start with AI.

For decades, architects have overprotected data, drawings, models, specs, not out of malice, but out of fear: fear of losing IP, fear of being copied, fear of liability. That's understandable. But it means owners don't receive the operational intelligence their buildings need.

Now, AI is simply exposing that longstanding issue. When researchers try to apply AI to the built environment, they discover there's no consistent data, no semantics, no identities, no systems to read. So they create synthetic buildings instead. That isn't an AI problem; it's an architecture and industry problem that's finally coming into view.

And this brings me back to your original point: should AI be allowed to reshape our industry? My answer is: only if we refuse to participate.

If we stay engaged, understand the tools, and bring our knowledge of buildings, systems, and human environments into the conversation, we can help steer AI toward outcomes that support owners, the public, and the profession.

If we withdraw, AI won't stop. It will simply move forward without our input.

That's why I believe the question isn't "friend or foe." It's: How do we, as architects, show up to guide this in a direction that reflects our values?

I would really like to hear your thoughts, and others', on what that guidance should look like. How do we balance legitimate concerns about IP, craft, and authorship with the responsibility to help shape tools that are already impacting the world around us?

This is exactly the kind of conversation the AIA should be having.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-09-2025 05:47 PM
From: Ian Toner, AIA
Subject: AI in the Studio: Friend or Foe?

I see a lot of posts that boil down to "AI is inevitable, so just give in". It's got me wondering--why does it need to be inevitable? It's not an earthquake or a sunrise. It's something created by people. It only exists if we choose to make it exist. It only takes over our industry if we decide to let it. It sounds really defeatist to say that we have to accept the consequences of AI just because some billionaires are running around saying it has to happen.

Sam Altman (CEO of OpenAI) was on Jimmy Fallon last night, saying that he couldn't imagine raising his child without AI to refer to (to quote him: "I cannot imagine having gone through figuring out how to raise a newborn without ChatGPT"). It feels like he's decided to give up an essential piece of his humanity--figuring stuff out. Or, having an interaction with another person who can share their perspective and wisdom as part of a real, human relationship.

Giving all our data to these guys for free, just so that they can develop tools that we get to pay a monthly subscription price for doesn't seem like a fair trade. Copyright law exists for a reason. IP takes time and skill to develop. It's not like they're asking, either--they're out there scraping the internet. Why not stand up to them?

I 100% realize I sound like a cranky old man with this. It's ok, I own it. I'm all for new tools and ideas, but this one feels pretty un-thought-through to me. It feels like there's a lot of FOMO here that's driving hasty decisions and not much thinking ahead. If a client came to you with this kind of rushed enthusiasm, wouldn't you tell them to slow down and spend a little time planning?

------------------------------
Ian Toner AIA
Toner Architects
Philadelphia PA

Original Message:
Sent: 12-09-2025 11:14 AM
From: Kimon Onuma, FAIA
Subject: AI in the Studio: Friend or Foe?

@Thesla Collier, Intl. Assoc. AIA I appreciate this conversation because it reflects where many architects are right now: curious, cautious, and trying to understand where AI fits into practice. But after presenting last week at NeurIPS in San Diego, the Conference on Neural Information Processing Systems, the world's largest and most influential AI research gathering, I can say this with confidence:

We are asking the wrong question.

While we debate whether AI is a "friend or foe" in the studio, NeurIPS brought together 22,000 researchers working on the systems that will shape the future of cities, mobility, infrastructure, and the built world.

Here is the critical insight:
NeurIPS is literally focused on building neural networks that describe the real world.
What they don't have, and what they openly said they need, is the neural network of buildings and cities themselves.

And that's where our profession is completely missing.

As an architect who has spent decades working on data, interoperability, and building intelligence, here's the hard truth:

The built environment is invisible to AI, not because the technology isn't ready, but because our profession refuses to expose the intelligence inside our buildings.

At Greenbuild this year, I heard the AIA Large Firm Roundtable (LFRT) say they strip data before handoff and avoid sharing models out of fear of losing IP. This is the same pattern we saw with CAD and BIM, and now the stakes are dramatically higher.

Meanwhile, at NeurIPS, I learned something the AEC industry has not yet absorbed:

Because architects don't share real operational intelligence, the AI industry is forced to use AI to create synthetic data about buildings and cities.

AI researchers are literally fabricating data for:

• building systems

• equipment behavior

• energy loads

• occupancy patterns

• maintenance events

• city infrastructure

• environmental interactions

Not because they want synthetic data, but because the real data is locked in PDFs, proprietary silos, and "means-and-methods" IP walls.

To AI researchers, our buildings appear:

• mute

• structureless

• disconnected

• uninterpretable

So AI does what it always does when data is missing:

It hallucinates the missing context and rebuilds our world from scratch.

This should concern us. It means the models shaping the future of cities do not contain architectural intelligence, because we have not provided it.

And here's the deeper issue:

If architects do not define machine-readable buildings and cities, AI will.

And it will do so without our expertise.

This is not about giving away creativity.
It's about giving away silence.

Shared, structured, machine-readable intelligence is one of the most sustainable materials we can contribute to the built world. If we continue treating data as proprietary instead of foundational, we hold back the very outcomes we claim to champion: climate performance, resilience, operational safety, great design, and long-term value for owners.

If you're interested in what I presented at NeurIPS, and what I saw there that directly affects our profession, I shared a summary here:

LinkedIn post:
https://www.linkedin.com/posts/kimononuma_neurips-urbanai-digitaltwins-activity-7404160879551799296-2eE-?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAErPBoBs61SLDRQoQjZHu1gU7qXr4VnM3I

AutomatedBuildings article:
https://www.automatedbuildings.com/2025/12/urban-ai-meets-the-built-environment-breaking-the-firewalls-that-keep-buildings-silent/

AI is enhancing creativity and replacing tasks, but it replaces only what we refuse to define.

Diagrams, specs, options, renderings, animations, 3D… AI can generate all of these synthetically because the real intelligence inside our buildings is locked behind our own walls.

Authorship remains where it always has:
in the architect's ability to give form, relationships, and meaning to the built environment, and now, to its data.

If we provide that foundation, AI becomes a collaborator. If we don't, AI will approximate the built world using synthetic models, and authorship simply shifts elsewhere.

AI is already reshaping our field.
The choice is whether we participate in shaping it or watch from the sidelines.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-04-2025 08:08 PM
From: Thesla Collier, Intl. Assoc. AIA
Subject: AI in the Studio: Friend or Foe?

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

------------------------------
Thesla Collier
HNTB
------------------------------

This thread about using AI to "read" drawings is interesting and related to the friend or foe topic:

https://www.linkedin.com/posts/caseyrutland_can-ai-really-read-your-building-plans-introducing-activity-7405216086641827840-nK-U?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAErPBoBs61SLDRQoQjZHu1gU7qXr4VnM3I

In this scenario, using an AI LLM to interpret and reverse-engineer architectural documents is a massive waste of resources if the source is BIM.

100%

I agree, trust is fundamental and humans must be in the loop. But I think trust comes less from the tool itself and more from clarity around what we're feeding it and what we expect in return.

If our workflows are still document-centric and ambiguous, it's hard to trust any system, AI or otherwise, because the inputs are opaque and the outcomes unpredictable. When building information is well defined, structured, and intentional, trust follows naturally.

In that sense, AI doesn't create trust on its own, it reflects the quality and discipline of the processes we already have.

 Kimon:

Ian, I appreciate your perspective, and I share many of the same concerns. I've been concerned about the trajectory of AI for a long time, not because I think it's inherently good or bad, but because it's powerful. And when something powerful emerges, the worst response is to ignore it. The best response is to understand it deeply enough to guide it.

That's how I approached AI from the beginning. The more I learned, the more I realized that our industry can't shape something we refuse to engage with.

You brought up billionaires pushing AI forward. I don't think we should accept their vision as inevitable. In fact, I believe it's our obligation as architects to help guide how AI interacts with the built world. 

It reminds me of raising teenagers: We don't ignore them because we don't like what they're doing; we stay involved, try to understand their thinking, and guide them toward better outcomes. If we disengage, they simply go their own way.

AI is the same. It's already here, and it will continue to evolve. If we step back, others will shape how it understands buildings, cities, and human environments, and that may not align with our values at all.

And this connects to something more profound: the lack of useful, machine-readable information inside buildings didn't start with AI.

For decades, architects have overprotected data, drawings, models, specs, not out of malice, but out of fear: fear of losing IP, fear of being copied, fear of liability. That's understandable. But it means owners don't receive the operational intelligence their buildings need.

Now, AI is simply exposing that longstanding issue. When researchers try to apply AI to the built environment, they discover there's no consistent data, no semantics, no identities, no systems to read. So they create synthetic buildings instead. That isn't an AI problem; it's an architecture and industry problem that's finally coming into view.

And this brings me back to your original point: should AI be allowed to reshape our industry? My answer is: only if we refuse to participate.

If we stay engaged, understand the tools, and bring our knowledge of buildings, systems, and human environments into the conversation, we can help steer AI toward outcomes that support owners, the public, and the profession.

If we withdraw, AI won't stop. It will simply move forward without our input.

That's why I believe the question isn't "friend or foe." It's: How do we, as architects, show up to guide this in a direction that reflects our values?

I would really like to hear your thoughts, and others', on what that guidance should look like. How do we balance legitimate concerns about IP, craft, and authorship with the responsibility to help shape tools that are already impacting the world around us?

This is exactly the kind of conversation the AIA should be having.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-09-2025 05:47 PM
From: Ian Toner, AIA
Subject: AI in the Studio: Friend or Foe?

I see a lot of posts that boil down to "AI is inevitable, so just give in". It's got me wondering--why does it need to be inevitable? It's not an earthquake or a sunrise. It's something created by people. It only exists if we choose to make it exist. It only takes over our industry if we decide to let it. It sounds really defeatist to say that we have to accept the consequences of AI just because some billionaires are running around saying it has to happen.

Sam Altman (CEO of OpenAI) was on Jimmy Fallon last night, saying that he couldn't imagine raising his child without AI to refer to (to quote him: "I cannot imagine having gone through figuring out how to raise a newborn without ChatGPT"). It feels like he's decided to give up an essential piece of his humanity--figuring stuff out. Or, having an interaction with another person who can share their perspective and wisdom as part of a real, human relationship.

Giving all our data to these guys for free, just so that they can develop tools that we get to pay a monthly subscription price for doesn't seem like a fair trade. Copyright law exists for a reason. IP takes time and skill to develop. It's not like they're asking, either--they're out there scraping the internet. Why not stand up to them?

I 100% realize I sound like a cranky old man with this. It's ok, I own it. I'm all for new tools and ideas, but this one feels pretty un-thought-through to me. It feels like there's a lot of FOMO here that's driving hasty decisions and not much thinking ahead. If a client came to you with this kind of rushed enthusiasm, wouldn't you tell them to slow down and spend a little time planning?

------------------------------
Ian Toner AIA
Toner Architects
Philadelphia PA

Original Message:
Sent: 12-09-2025 11:14 AM
From: Kimon Onuma, FAIA
Subject: AI in the Studio: Friend or Foe?

@Thesla Collier, Intl. Assoc. AIA I appreciate this conversation because it reflects where many architects are right now: curious, cautious, and trying to understand where AI fits into practice. But after presenting last week at NeurIPS in San Diego, the Conference on Neural Information Processing Systems, the world's largest and most influential AI research gathering, I can say this with confidence:

We are asking the wrong question.

While we debate whether AI is a "friend or foe" in the studio, NeurIPS brought together 22,000 researchers working on the systems that will shape the future of cities, mobility, infrastructure, and the built world.

Here is the critical insight:
NeurIPS is literally focused on building neural networks that describe the real world.
What they don't have, and what they openly said they need, is the neural network of buildings and cities themselves.

And that's where our profession is completely missing.

As an architect who has spent decades working on data, interoperability, and building intelligence, here's the hard truth:

The built environment is invisible to AI, not because the technology isn't ready, but because our profession refuses to expose the intelligence inside our buildings.

At Greenbuild this year, I heard the AIA Large Firm Roundtable (LFRT) say they strip data before handoff and avoid sharing models out of fear of losing IP. This is the same pattern we saw with CAD and BIM, and now the stakes are dramatically higher.

Meanwhile, at NeurIPS, I learned something the AEC industry has not yet absorbed:

Because architects don't share real operational intelligence, the AI industry is forced to use AI to create synthetic data about buildings and cities.

AI researchers are literally fabricating data for:

• building systems

• equipment behavior

• energy loads

• occupancy patterns

• maintenance events

• city infrastructure

• environmental interactions

Not because they want synthetic data, but because the real data is locked in PDFs, proprietary silos, and "means-and-methods" IP walls.

To AI researchers, our buildings appear:

• mute

• structureless

• disconnected

• uninterpretable

So AI does what it always does when data is missing:

It hallucinates the missing context and rebuilds our world from scratch.

This should concern us. It means the models shaping the future of cities do not contain architectural intelligence, because we have not provided it.

And here's the deeper issue:

If architects do not define machine-readable buildings and cities, AI will.

And it will do so without our expertise.

This is not about giving away creativity.
It's about giving away silence.

Shared, structured, machine-readable intelligence is one of the most sustainable materials we can contribute to the built world. If we continue treating data as proprietary instead of foundational, we hold back the very outcomes we claim to champion: climate performance, resilience, operational safety, great design, and long-term value for owners.

If you're interested in what I presented at NeurIPS, and what I saw there that directly affects our profession, I shared a summary here:

LinkedIn post:
https://www.linkedin.com/posts/kimononuma_neurips-urbanai-digitaltwins-activity-7404160879551799296-2eE-?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAErPBoBs61SLDRQoQjZHu1gU7qXr4VnM3I

AutomatedBuildings article:
https://www.automatedbuildings.com/2025/12/urban-ai-meets-the-built-environment-breaking-the-firewalls-that-keep-buildings-silent/

AI is enhancing creativity and replacing tasks, but it replaces only what we refuse to define.

Diagrams, specs, options, renderings, animations, 3D… AI can generate all of these synthetically because the real intelligence inside our buildings is locked behind our own walls.

Authorship remains where it always has:
in the architect's ability to give form, relationships, and meaning to the built environment, and now, to its data.

If we provide that foundation, AI becomes a collaborator. If we don't, AI will approximate the built world using synthetic models, and authorship simply shifts elsewhere.

AI is already reshaping our field.
The choice is whether we participate in shaping it or watch from the sidelines.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-04-2025 08:08 PM
From: Thesla Collier, Intl. Assoc. AIA
Subject: AI in the Studio: Friend or Foe?

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

------------------------------
Thesla Collier
HNTB
------------------------------

I see this as a temporary plateau. AI will certainly learn how to "read" plans. It will then do what it does best- it will improve and instruct each succeeding generation of software. I have no doubt that, in a years time, it will be able to accurately count doors and windows in a snap.

This thread about using AI to "read" drawings is interesting and related to the friend or foe topic:

https://www.linkedin.com/posts/caseyrutland_can-ai-really-read-your-building-plans-introducing-activity-7405216086641827840-nK-U?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAErPBoBs61SLDRQoQjZHu1gU7qXr4VnM3I

In this scenario, using an AI LLM to interpret and reverse-engineer architectural documents is a massive waste of resources if the source is BIM.

100%

I agree, trust is fundamental and humans must be in the loop. But I think trust comes less from the tool itself and more from clarity around what we're feeding it and what we expect in return.

If our workflows are still document-centric and ambiguous, it's hard to trust any system, AI or otherwise, because the inputs are opaque and the outcomes unpredictable. When building information is well defined, structured, and intentional, trust follows naturally.

In that sense, AI doesn't create trust on its own, it reflects the quality and discipline of the processes we already have.

 Kimon:

Ian, I appreciate your perspective, and I share many of the same concerns. I've been concerned about the trajectory of AI for a long time, not because I think it's inherently good or bad, but because it's powerful. And when something powerful emerges, the worst response is to ignore it. The best response is to understand it deeply enough to guide it.

That's how I approached AI from the beginning. The more I learned, the more I realized that our industry can't shape something we refuse to engage with.

You brought up billionaires pushing AI forward. I don't think we should accept their vision as inevitable. In fact, I believe it's our obligation as architects to help guide how AI interacts with the built world. 

It reminds me of raising teenagers: We don't ignore them because we don't like what they're doing; we stay involved, try to understand their thinking, and guide them toward better outcomes. If we disengage, they simply go their own way.

AI is the same. It's already here, and it will continue to evolve. If we step back, others will shape how it understands buildings, cities, and human environments, and that may not align with our values at all.

And this connects to something more profound: the lack of useful, machine-readable information inside buildings didn't start with AI.

For decades, architects have overprotected data, drawings, models, specs, not out of malice, but out of fear: fear of losing IP, fear of being copied, fear of liability. That's understandable. But it means owners don't receive the operational intelligence their buildings need.

Now, AI is simply exposing that longstanding issue. When researchers try to apply AI to the built environment, they discover there's no consistent data, no semantics, no identities, no systems to read. So they create synthetic buildings instead. That isn't an AI problem; it's an architecture and industry problem that's finally coming into view.

And this brings me back to your original point: should AI be allowed to reshape our industry? My answer is: only if we refuse to participate.

If we stay engaged, understand the tools, and bring our knowledge of buildings, systems, and human environments into the conversation, we can help steer AI toward outcomes that support owners, the public, and the profession.

If we withdraw, AI won't stop. It will simply move forward without our input.

That's why I believe the question isn't "friend or foe." It's: How do we, as architects, show up to guide this in a direction that reflects our values?

I would really like to hear your thoughts, and others', on what that guidance should look like. How do we balance legitimate concerns about IP, craft, and authorship with the responsibility to help shape tools that are already impacting the world around us?

This is exactly the kind of conversation the AIA should be having.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-09-2025 05:47 PM
From: Ian Toner, AIA
Subject: AI in the Studio: Friend or Foe?

I see a lot of posts that boil down to "AI is inevitable, so just give in". It's got me wondering--why does it need to be inevitable? It's not an earthquake or a sunrise. It's something created by people. It only exists if we choose to make it exist. It only takes over our industry if we decide to let it. It sounds really defeatist to say that we have to accept the consequences of AI just because some billionaires are running around saying it has to happen.

Sam Altman (CEO of OpenAI) was on Jimmy Fallon last night, saying that he couldn't imagine raising his child without AI to refer to (to quote him: "I cannot imagine having gone through figuring out how to raise a newborn without ChatGPT"). It feels like he's decided to give up an essential piece of his humanity--figuring stuff out. Or, having an interaction with another person who can share their perspective and wisdom as part of a real, human relationship.

Giving all our data to these guys for free, just so that they can develop tools that we get to pay a monthly subscription price for doesn't seem like a fair trade. Copyright law exists for a reason. IP takes time and skill to develop. It's not like they're asking, either--they're out there scraping the internet. Why not stand up to them?

I 100% realize I sound like a cranky old man with this. It's ok, I own it. I'm all for new tools and ideas, but this one feels pretty un-thought-through to me. It feels like there's a lot of FOMO here that's driving hasty decisions and not much thinking ahead. If a client came to you with this kind of rushed enthusiasm, wouldn't you tell them to slow down and spend a little time planning?

------------------------------
Ian Toner AIA
Toner Architects
Philadelphia PA

Original Message:
Sent: 12-09-2025 11:14 AM
From: Kimon Onuma, FAIA
Subject: AI in the Studio: Friend or Foe?

@Thesla Collier, Intl. Assoc. AIA I appreciate this conversation because it reflects where many architects are right now: curious, cautious, and trying to understand where AI fits into practice. But after presenting last week at NeurIPS in San Diego, the Conference on Neural Information Processing Systems, the world's largest and most influential AI research gathering, I can say this with confidence:

We are asking the wrong question.

While we debate whether AI is a "friend or foe" in the studio, NeurIPS brought together 22,000 researchers working on the systems that will shape the future of cities, mobility, infrastructure, and the built world.

Here is the critical insight:
NeurIPS is literally focused on building neural networks that describe the real world.
What they don't have, and what they openly said they need, is the neural network of buildings and cities themselves.

And that's where our profession is completely missing.

As an architect who has spent decades working on data, interoperability, and building intelligence, here's the hard truth:

The built environment is invisible to AI, not because the technology isn't ready, but because our profession refuses to expose the intelligence inside our buildings.

At Greenbuild this year, I heard the AIA Large Firm Roundtable (LFRT) say they strip data before handoff and avoid sharing models out of fear of losing IP. This is the same pattern we saw with CAD and BIM, and now the stakes are dramatically higher.

Meanwhile, at NeurIPS, I learned something the AEC industry has not yet absorbed:

Because architects don't share real operational intelligence, the AI industry is forced to use AI to create synthetic data about buildings and cities.

AI researchers are literally fabricating data for:

• building systems

• equipment behavior

• energy loads

• occupancy patterns

• maintenance events

• city infrastructure

• environmental interactions

Not because they want synthetic data, but because the real data is locked in PDFs, proprietary silos, and "means-and-methods" IP walls.

To AI researchers, our buildings appear:

• mute

• structureless

• disconnected

• uninterpretable

So AI does what it always does when data is missing:

It hallucinates the missing context and rebuilds our world from scratch.

This should concern us. It means the models shaping the future of cities do not contain architectural intelligence, because we have not provided it.

And here's the deeper issue:

If architects do not define machine-readable buildings and cities, AI will.

And it will do so without our expertise.

This is not about giving away creativity.
It's about giving away silence.

Shared, structured, machine-readable intelligence is one of the most sustainable materials we can contribute to the built world. If we continue treating data as proprietary instead of foundational, we hold back the very outcomes we claim to champion: climate performance, resilience, operational safety, great design, and long-term value for owners.

If you're interested in what I presented at NeurIPS, and what I saw there that directly affects our profession, I shared a summary here:

LinkedIn post:
https://www.linkedin.com/posts/kimononuma_neurips-urbanai-digitaltwins-activity-7404160879551799296-2eE-?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAErPBoBs61SLDRQoQjZHu1gU7qXr4VnM3I

AutomatedBuildings article:
https://www.automatedbuildings.com/2025/12/urban-ai-meets-the-built-environment-breaking-the-firewalls-that-keep-buildings-silent/

AI is enhancing creativity and replacing tasks, but it replaces only what we refuse to define.

Diagrams, specs, options, renderings, animations, 3D… AI can generate all of these synthetically because the real intelligence inside our buildings is locked behind our own walls.

Authorship remains where it always has:
in the architect's ability to give form, relationships, and meaning to the built environment, and now, to its data.

If we provide that foundation, AI becomes a collaborator. If we don't, AI will approximate the built world using synthetic models, and authorship simply shifts elsewhere.

AI is already reshaping our field.
The choice is whether we participate in shaping it or watch from the sidelines.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-04-2025 08:08 PM
From: Thesla Collier, Intl. Assoc. AIA
Subject: AI in the Studio: Friend or Foe?

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

------------------------------
Thesla Collier
HNTB
------------------------------

Steven: the tools are getting more accurate each day. What matters is the quality of the underlying graphic. For example if there are symbols, text, dashed lines sitting on top of doors and windows even humans would have a hard time! One tool we have been using is https://wisebim.app/  there are several others.

I see this as a temporary plateau. AI will certainly learn how to "read" plans. It will then do what it does best- it will improve and instruct each succeeding generation of software. I have no doubt that, in a years time, it will be able to accurately count doors and windows in a snap.

This thread about using AI to "read" drawings is interesting and related to the friend or foe topic:

https://www.linkedin.com/posts/caseyrutland_can-ai-really-read-your-building-plans-introducing-activity-7405216086641827840-nK-U?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAErPBoBs61SLDRQoQjZHu1gU7qXr4VnM3I

In this scenario, using an AI LLM to interpret and reverse-engineer architectural documents is a massive waste of resources if the source is BIM.

100%

I agree, trust is fundamental and humans must be in the loop. But I think trust comes less from the tool itself and more from clarity around what we're feeding it and what we expect in return.

If our workflows are still document-centric and ambiguous, it's hard to trust any system, AI or otherwise, because the inputs are opaque and the outcomes unpredictable. When building information is well defined, structured, and intentional, trust follows naturally.

In that sense, AI doesn't create trust on its own, it reflects the quality and discipline of the processes we already have.

 Kimon:

Ian, I appreciate your perspective, and I share many of the same concerns. I've been concerned about the trajectory of AI for a long time, not because I think it's inherently good or bad, but because it's powerful. And when something powerful emerges, the worst response is to ignore it. The best response is to understand it deeply enough to guide it.

That's how I approached AI from the beginning. The more I learned, the more I realized that our industry can't shape something we refuse to engage with.

You brought up billionaires pushing AI forward. I don't think we should accept their vision as inevitable. In fact, I believe it's our obligation as architects to help guide how AI interacts with the built world. 

It reminds me of raising teenagers: We don't ignore them because we don't like what they're doing; we stay involved, try to understand their thinking, and guide them toward better outcomes. If we disengage, they simply go their own way.

AI is the same. It's already here, and it will continue to evolve. If we step back, others will shape how it understands buildings, cities, and human environments, and that may not align with our values at all.

And this connects to something more profound: the lack of useful, machine-readable information inside buildings didn't start with AI.

For decades, architects have overprotected data, drawings, models, specs, not out of malice, but out of fear: fear of losing IP, fear of being copied, fear of liability. That's understandable. But it means owners don't receive the operational intelligence their buildings need.

Now, AI is simply exposing that longstanding issue. When researchers try to apply AI to the built environment, they discover there's no consistent data, no semantics, no identities, no systems to read. So they create synthetic buildings instead. That isn't an AI problem; it's an architecture and industry problem that's finally coming into view.

And this brings me back to your original point: should AI be allowed to reshape our industry? My answer is: only if we refuse to participate.

If we stay engaged, understand the tools, and bring our knowledge of buildings, systems, and human environments into the conversation, we can help steer AI toward outcomes that support owners, the public, and the profession.

If we withdraw, AI won't stop. It will simply move forward without our input.

That's why I believe the question isn't "friend or foe." It's: How do we, as architects, show up to guide this in a direction that reflects our values?

I would really like to hear your thoughts, and others', on what that guidance should look like. How do we balance legitimate concerns about IP, craft, and authorship with the responsibility to help shape tools that are already impacting the world around us?

This is exactly the kind of conversation the AIA should be having.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-09-2025 05:47 PM
From: Ian Toner, AIA
Subject: AI in the Studio: Friend or Foe?

I see a lot of posts that boil down to "AI is inevitable, so just give in". It's got me wondering--why does it need to be inevitable? It's not an earthquake or a sunrise. It's something created by people. It only exists if we choose to make it exist. It only takes over our industry if we decide to let it. It sounds really defeatist to say that we have to accept the consequences of AI just because some billionaires are running around saying it has to happen.

Sam Altman (CEO of OpenAI) was on Jimmy Fallon last night, saying that he couldn't imagine raising his child without AI to refer to (to quote him: "I cannot imagine having gone through figuring out how to raise a newborn without ChatGPT"). It feels like he's decided to give up an essential piece of his humanity--figuring stuff out. Or, having an interaction with another person who can share their perspective and wisdom as part of a real, human relationship.

Giving all our data to these guys for free, just so that they can develop tools that we get to pay a monthly subscription price for doesn't seem like a fair trade. Copyright law exists for a reason. IP takes time and skill to develop. It's not like they're asking, either--they're out there scraping the internet. Why not stand up to them?

I 100% realize I sound like a cranky old man with this. It's ok, I own it. I'm all for new tools and ideas, but this one feels pretty un-thought-through to me. It feels like there's a lot of FOMO here that's driving hasty decisions and not much thinking ahead. If a client came to you with this kind of rushed enthusiasm, wouldn't you tell them to slow down and spend a little time planning?

------------------------------
Ian Toner AIA
Toner Architects
Philadelphia PA

Original Message:
Sent: 12-09-2025 11:14 AM
From: Kimon Onuma, FAIA
Subject: AI in the Studio: Friend or Foe?

@Thesla Collier, Intl. Assoc. AIA I appreciate this conversation because it reflects where many architects are right now: curious, cautious, and trying to understand where AI fits into practice. But after presenting last week at NeurIPS in San Diego, the Conference on Neural Information Processing Systems, the world's largest and most influential AI research gathering, I can say this with confidence:

We are asking the wrong question.

While we debate whether AI is a "friend or foe" in the studio, NeurIPS brought together 22,000 researchers working on the systems that will shape the future of cities, mobility, infrastructure, and the built world.

Here is the critical insight:
NeurIPS is literally focused on building neural networks that describe the real world.
What they don't have, and what they openly said they need, is the neural network of buildings and cities themselves.

And that's where our profession is completely missing.

As an architect who has spent decades working on data, interoperability, and building intelligence, here's the hard truth:

The built environment is invisible to AI, not because the technology isn't ready, but because our profession refuses to expose the intelligence inside our buildings.

At Greenbuild this year, I heard the AIA Large Firm Roundtable (LFRT) say they strip data before handoff and avoid sharing models out of fear of losing IP. This is the same pattern we saw with CAD and BIM, and now the stakes are dramatically higher.

Meanwhile, at NeurIPS, I learned something the AEC industry has not yet absorbed:

Because architects don't share real operational intelligence, the AI industry is forced to use AI to create synthetic data about buildings and cities.

AI researchers are literally fabricating data for:

• building systems

• equipment behavior

• energy loads

• occupancy patterns

• maintenance events

• city infrastructure

• environmental interactions

Not because they want synthetic data, but because the real data is locked in PDFs, proprietary silos, and "means-and-methods" IP walls.

To AI researchers, our buildings appear:

• mute

• structureless

• disconnected

• uninterpretable

So AI does what it always does when data is missing:

It hallucinates the missing context and rebuilds our world from scratch.

This should concern us. It means the models shaping the future of cities do not contain architectural intelligence, because we have not provided it.

And here's the deeper issue:

If architects do not define machine-readable buildings and cities, AI will.

And it will do so without our expertise.

This is not about giving away creativity.
It's about giving away silence.

Shared, structured, machine-readable intelligence is one of the most sustainable materials we can contribute to the built world. If we continue treating data as proprietary instead of foundational, we hold back the very outcomes we claim to champion: climate performance, resilience, operational safety, great design, and long-term value for owners.

If you're interested in what I presented at NeurIPS, and what I saw there that directly affects our profession, I shared a summary here:

LinkedIn post:
https://www.linkedin.com/posts/kimononuma_neurips-urbanai-digitaltwins-activity-7404160879551799296-2eE-?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAErPBoBs61SLDRQoQjZHu1gU7qXr4VnM3I

AutomatedBuildings article:
https://www.automatedbuildings.com/2025/12/urban-ai-meets-the-built-environment-breaking-the-firewalls-that-keep-buildings-silent/

AI is enhancing creativity and replacing tasks, but it replaces only what we refuse to define.

Diagrams, specs, options, renderings, animations, 3D… AI can generate all of these synthetically because the real intelligence inside our buildings is locked behind our own walls.

Authorship remains where it always has:
in the architect's ability to give form, relationships, and meaning to the built environment, and now, to its data.

If we provide that foundation, AI becomes a collaborator. If we don't, AI will approximate the built world using synthetic models, and authorship simply shifts elsewhere.

AI is already reshaping our field.
The choice is whether we participate in shaping it or watch from the sidelines.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-04-2025 08:08 PM
From: Thesla Collier, Intl. Assoc. AIA
Subject: AI in the Studio: Friend or Foe?

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

------------------------------
Thesla Collier
HNTB
------------------------------

Steven: the tools are getting more accurate each day. What matters is the quality of the underlying graphic. For example if there are symbols, text, dashed lines sitting on top of doors and windows even humans would have a hard time! One tool we have been using is https://wisebim.app/  there are several others.

I see this as a temporary plateau. AI will certainly learn how to "read" plans. It will then do what it does best- it will improve and instruct each succeeding generation of software. I have no doubt that, in a years time, it will be able to accurately count doors and windows in a snap.

This thread about using AI to "read" drawings is interesting and related to the friend or foe topic:

https://www.linkedin.com/posts/caseyrutland_can-ai-really-read-your-building-plans-introducing-activity-7405216086641827840-nK-U?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAErPBoBs61SLDRQoQjZHu1gU7qXr4VnM3I

In this scenario, using an AI LLM to interpret and reverse-engineer architectural documents is a massive waste of resources if the source is BIM.

100%

I agree, trust is fundamental and humans must be in the loop. But I think trust comes less from the tool itself and more from clarity around what we're feeding it and what we expect in return.

If our workflows are still document-centric and ambiguous, it's hard to trust any system, AI or otherwise, because the inputs are opaque and the outcomes unpredictable. When building information is well defined, structured, and intentional, trust follows naturally.

In that sense, AI doesn't create trust on its own, it reflects the quality and discipline of the processes we already have.

 Kimon:

Ian, I appreciate your perspective, and I share many of the same concerns. I've been concerned about the trajectory of AI for a long time, not because I think it's inherently good or bad, but because it's powerful. And when something powerful emerges, the worst response is to ignore it. The best response is to understand it deeply enough to guide it.

That's how I approached AI from the beginning. The more I learned, the more I realized that our industry can't shape something we refuse to engage with.

You brought up billionaires pushing AI forward. I don't think we should accept their vision as inevitable. In fact, I believe it's our obligation as architects to help guide how AI interacts with the built world. 

It reminds me of raising teenagers: We don't ignore them because we don't like what they're doing; we stay involved, try to understand their thinking, and guide them toward better outcomes. If we disengage, they simply go their own way.

AI is the same. It's already here, and it will continue to evolve. If we step back, others will shape how it understands buildings, cities, and human environments, and that may not align with our values at all.

And this connects to something more profound: the lack of useful, machine-readable information inside buildings didn't start with AI.

For decades, architects have overprotected data, drawings, models, specs, not out of malice, but out of fear: fear of losing IP, fear of being copied, fear of liability. That's understandable. But it means owners don't receive the operational intelligence their buildings need.

Now, AI is simply exposing that longstanding issue. When researchers try to apply AI to the built environment, they discover there's no consistent data, no semantics, no identities, no systems to read. So they create synthetic buildings instead. That isn't an AI problem; it's an architecture and industry problem that's finally coming into view.

And this brings me back to your original point: should AI be allowed to reshape our industry? My answer is: only if we refuse to participate.

If we stay engaged, understand the tools, and bring our knowledge of buildings, systems, and human environments into the conversation, we can help steer AI toward outcomes that support owners, the public, and the profession.

If we withdraw, AI won't stop. It will simply move forward without our input.

That's why I believe the question isn't "friend or foe." It's: How do we, as architects, show up to guide this in a direction that reflects our values?

I would really like to hear your thoughts, and others', on what that guidance should look like. How do we balance legitimate concerns about IP, craft, and authorship with the responsibility to help shape tools that are already impacting the world around us?

This is exactly the kind of conversation the AIA should be having.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-09-2025 05:47 PM
From: Ian Toner, AIA
Subject: AI in the Studio: Friend or Foe?

I see a lot of posts that boil down to "AI is inevitable, so just give in". It's got me wondering--why does it need to be inevitable? It's not an earthquake or a sunrise. It's something created by people. It only exists if we choose to make it exist. It only takes over our industry if we decide to let it. It sounds really defeatist to say that we have to accept the consequences of AI just because some billionaires are running around saying it has to happen.

Sam Altman (CEO of OpenAI) was on Jimmy Fallon last night, saying that he couldn't imagine raising his child without AI to refer to (to quote him: "I cannot imagine having gone through figuring out how to raise a newborn without ChatGPT"). It feels like he's decided to give up an essential piece of his humanity--figuring stuff out. Or, having an interaction with another person who can share their perspective and wisdom as part of a real, human relationship.

Giving all our data to these guys for free, just so that they can develop tools that we get to pay a monthly subscription price for doesn't seem like a fair trade. Copyright law exists for a reason. IP takes time and skill to develop. It's not like they're asking, either--they're out there scraping the internet. Why not stand up to them?

I 100% realize I sound like a cranky old man with this. It's ok, I own it. I'm all for new tools and ideas, but this one feels pretty un-thought-through to me. It feels like there's a lot of FOMO here that's driving hasty decisions and not much thinking ahead. If a client came to you with this kind of rushed enthusiasm, wouldn't you tell them to slow down and spend a little time planning?

------------------------------
Ian Toner AIA
Toner Architects
Philadelphia PA

Original Message:
Sent: 12-09-2025 11:14 AM
From: Kimon Onuma, FAIA
Subject: AI in the Studio: Friend or Foe?

@Thesla Collier, Intl. Assoc. AIA I appreciate this conversation because it reflects where many architects are right now: curious, cautious, and trying to understand where AI fits into practice. But after presenting last week at NeurIPS in San Diego, the Conference on Neural Information Processing Systems, the world's largest and most influential AI research gathering, I can say this with confidence:

We are asking the wrong question.

While we debate whether AI is a "friend or foe" in the studio, NeurIPS brought together 22,000 researchers working on the systems that will shape the future of cities, mobility, infrastructure, and the built world.

Here is the critical insight:
NeurIPS is literally focused on building neural networks that describe the real world.
What they don't have, and what they openly said they need, is the neural network of buildings and cities themselves.

And that's where our profession is completely missing.

As an architect who has spent decades working on data, interoperability, and building intelligence, here's the hard truth:

The built environment is invisible to AI, not because the technology isn't ready, but because our profession refuses to expose the intelligence inside our buildings.

At Greenbuild this year, I heard the AIA Large Firm Roundtable (LFRT) say they strip data before handoff and avoid sharing models out of fear of losing IP. This is the same pattern we saw with CAD and BIM, and now the stakes are dramatically higher.

Meanwhile, at NeurIPS, I learned something the AEC industry has not yet absorbed:

Because architects don't share real operational intelligence, the AI industry is forced to use AI to create synthetic data about buildings and cities.

AI researchers are literally fabricating data for:

• building systems

• equipment behavior

• energy loads

• occupancy patterns

• maintenance events

• city infrastructure

• environmental interactions

Not because they want synthetic data, but because the real data is locked in PDFs, proprietary silos, and "means-and-methods" IP walls.

To AI researchers, our buildings appear:

• mute

• structureless

• disconnected

• uninterpretable

So AI does what it always does when data is missing:

It hallucinates the missing context and rebuilds our world from scratch.

This should concern us. It means the models shaping the future of cities do not contain architectural intelligence, because we have not provided it.

And here's the deeper issue:

If architects do not define machine-readable buildings and cities, AI will.

And it will do so without our expertise.

This is not about giving away creativity.
It's about giving away silence.

Shared, structured, machine-readable intelligence is one of the most sustainable materials we can contribute to the built world. If we continue treating data as proprietary instead of foundational, we hold back the very outcomes we claim to champion: climate performance, resilience, operational safety, great design, and long-term value for owners.

If you're interested in what I presented at NeurIPS, and what I saw there that directly affects our profession, I shared a summary here:

LinkedIn post:
https://www.linkedin.com/posts/kimononuma_neurips-urbanai-digitaltwins-activity-7404160879551799296-2eE-?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAErPBoBs61SLDRQoQjZHu1gU7qXr4VnM3I

AutomatedBuildings article:
https://www.automatedbuildings.com/2025/12/urban-ai-meets-the-built-environment-breaking-the-firewalls-that-keep-buildings-silent/

AI is enhancing creativity and replacing tasks, but it replaces only what we refuse to define.

Diagrams, specs, options, renderings, animations, 3D… AI can generate all of these synthetically because the real intelligence inside our buildings is locked behind our own walls.

Authorship remains where it always has:
in the architect's ability to give form, relationships, and meaning to the built environment, and now, to its data.

If we provide that foundation, AI becomes a collaborator. If we don't, AI will approximate the built world using synthetic models, and authorship simply shifts elsewhere.

AI is already reshaping our field.
The choice is whether we participate in shaping it or watch from the sidelines.

------------------------------
Kimon Onuma FAIA
Onuma, Inc.
Pasadena CA

Original Message:
Sent: 12-04-2025 08:08 PM
From: Thesla Collier, Intl. Assoc. AIA
Subject: AI in the Studio: Friend or Foe?

 Is AI enhancing creativity or replacing it?

With AI tools generating diagrams, specs, and even design options, where do you draw the line between assistance and authorship?

Share your stance and one example.

------------------------------
Thesla Collier
HNTB
------------------------------