A primary document, counted

Who signed Pacing the Frontier

In July 2026, employees of the companies building the most advanced AI models signed a one-sentence request to the U.S. government. By 23 September 2026 the statement site printed 1,386 signatories. This page publishes all of them — every name, job title and employer the statement site shows, plus all 96 of the public comments, quoted in full.

This is the source list itself, not a summary of who signed. The list was read off pacingthefrontier.com on 23 September 2026 and reproduced as published. Names are in the order the site lists them.

What the phrase means. “Pacing” was chosen over “slowing” or “pausing” on purpose, and the difference matters — the glossary entry explains it, and says plainly that the people proposing to pace the frontier are the people building it.

What it actually asks

“We request that the U.S. government support an international effort to develop the technical and governance tools needed to deliberately pace the frontier of automated AI development.”
— 1,386 employees of frontier AI companies, published 28 July 2026

That is the whole ask. The surrounding text argues that the leading companies may be close to automating AI research, that each company and country is under competitive pressure not to slow down on its own, and that the world currently lacks the tools to pace progress across the industry even if it wanted to. It does not request a pause, and it commits no company to anything.

The number has moved

People still cite “more than 1,000” or “over 1,200” because the statement is still open for signature and every report froze it at a different moment. If you quote a figure, quote the date with it.

1,13428 July 2026
The day the statement was published. This figure comes from an archived copy of the statement site taken that evening.
Internet Archive capture of pacingthefrontier.com
more than 1,20029 July 2026
Fortune's report, the day after publication. It named Dario Amodei, Jakub Pachocki, Shengjia Zhao and Anca Dragan among the signatories.
Fortune
1,38623 September 2026
The count printed on the statement site when this page was built. Every name, job title and comment below was read straight off that page on this date.
pacingthefrontier.com

Who signed, by employer

1,094 signatories are named. 292 are published as “Anonymous” with only an employer. Four companies account for almost everything: the rest of the list is a long tail of one, two and three.

Employer, as the site prints it	Signatories	Of those, named
Anthropic	587	452
OpenAI	394	335
Google	260	188
Meta	95	76
Thinking Machines	11	11
Amazon	7	6
Microsoft	6	6
Prime Intellect	4	4
Inherent	3	3
Standard Intelligence	3	3
Humans&	3	2
Hugging Face	2	2
EleutherAI	2	2
Cursor	2	0
Mistral	2	0
Safe Superintelligence	1	1
Core Automation	1	1
Cohere	1	1
Cognition	1	1
Imbue	1	0

Signatory timeline

All 1,386 entries, grouped by the earliest dated snapshot where each can be verified. The source publishes no individual signing timestamps, so “By 23 September” means only that the person appeared sometime after the launch-day snapshot and no later than this site’s capture.

What can be dated: 1,132 current entries already appear in the 28 July archive; 254 first appear in the later 23 September capture. Two launch-day names no longer appear in the current list. Public statements are shown only where the statement site explicitly connects the quote to that person.

Named onlyPublic statement only

Showing 60 of 1386 matching signatories.

28 July 2026Earliest verified
John Schulman
Chief Scientist, Thinking Machines
“Signed because this statement helps establish common knowledge about the possible need for coordination mechanisms as automated AI research accelerates progress. I'd also like to see labs start designing these mechanisms voluntarily, even before the USG gets involved.”
28 July 2026Earliest verified
Jakub Pachocki
Chief Scientist, OpenAI
No public statement attached.
28 July 2026Earliest verified
Jared Kaplan
Co-Founder and Chief Science Officer, Anthropic
No public statement attached.
28 July 2026Earliest verified
Shengjia Zhao
Chief Scientist, Meta
“AI is progressing at a rate that our society might not be ready for. Frontier labs are very close to AI that can exceed even the best people on almost every metric of intelligence. This will lead to unprecedented social and safety risks. To ensure a positive future we need to develop AI in a way that is driven by responsibility and thoughtfulness.”
By 23 September 2026Earliest verified
Shane Legg
Co-Founder & Chief AGI Scientist, Google
No public statement attached.
By 23 September 2026Earliest verified
Ilya Sutskever
CEO, Safe Superintelligence
“Future AI will be extraordinarily powerful compared to anything that exists today, and dealing with this future power will require unprecedented measures, such as the ones described here. The problem statement is real. This works only if it is done internationally, and it has to be done well: a bad implementation can make things worse.”
28 July 2026Earliest verified
Mark Chen
Chief Research Officer, OpenAI
No public statement attached.
28 July 2026Earliest verified
Jasjeet Sekhon
Chief Strategy Officer, Google
“We have found a way to turn energy into compute, and compute into intelligence. The benefits will be enormous, from curing diseases to understanding the cosmos. We can capture the benefits of the coming intelligence explosion while managing its risks, but only if we build the tools to pace the frontier of the riskiest capabilities before we need them, so we protect people and keep the social trust that innovation depends on. I believe smart technical and governance tools will be needed to sustain rapid innovation, vigorous competition, and robust safety.”
28 July 2026Earliest verified
Dario Amodei
CEO, Anthropic
No public statement attached.
28 July 2026Earliest verified
Jack Clark
Co-Founder and Head of Public Benefit, Anthropic
No public statement attached.
28 July 2026Earliest verified
Anca Dragan
VP, AI Safety & Alignment, Google
No public statement attached.
28 July 2026Earliest verified
Wojciech Zaremba
Head of AI Resilience, OpenAI
No public statement attached.
28 July 2026Earliest verified
Dawn Song
VP, AI Research, Meta
No public statement attached.
28 July 2026Earliest verified
Chris Olah
Co-Founder & Interpretability Research Lead, Anthropic
No public statement attached.
28 July 2026Earliest verified
Laura Weidinger
Staff Research Scientist, Google
“If it's true that AI will be as transformational as electricity, we better be intentional about how it is introduced to the world. A mindless race without assurances on governance, safety or the technology serving the public interest stands in the way of such intentionality.”
28 July 2026Earliest verified
Benjamin Mann
Co-Founder, Anthropic
No public statement attached.
28 July 2026Earliest verified
Stephanie Chan
Staff Research Scientist, Google
“I have worked in AI for a decade. Every few months in the last years, I have found myself surprised again and again at how rapidly the technology advances, no matter how often I update my predictions. AI may end up progressing far more quickly than societies and states can adapt. Thus, many prominent leaders in AI have publicly supported the ability to slow down, including DeepMind CEO Demis Hassabis. At the same time, coordination is hampered by multiple factors, including competitive dynamics and antitrust regulations. I signed this statement (in a personal capacity) because I support greater *optionality* in the range of actions available, to ensure that states and societies around the world have time to thoughtfully prepare and adapt to this powerful technology.”
28 July 2026Earliest verified
Julian Schrittwieser
Anthropic
No public statement attached.
28 July 2026Earliest verified
Summer Yue
Director of Alignment and Risk, Meta
No public statement attached.
28 July 2026Earliest verified
Jan Leike
Anthropic
No public statement attached.
28 July 2026Earliest verified
Boaz Barak
OpenAI
No public statement attached.
28 July 2026Earliest verified
Mike Krieger
Member of Technical Staff, Anthropic
No public statement attached.
28 July 2026Earliest verified
Mary Phuong
Senior Research Scientist, Google
No public statement attached.
28 July 2026Earliest verified
Shantanu Jain
Member of Technical Staff, OpenAI
“In the span of four years, we have gone from people having their first experience of an Al that could understand language, to Al that is performing superhuman feats of software engineering, making frontier math breakthroughs, and accelerating scientific research. Progress has shown no signs of slowing down and by default it is in the interest of individual corporations and countries to push the accelerator all the way down, even if it is in humanity's best interest for the future to not arrive all at once. Society has spent longer than we can remember developing the laws, institutions and technology to govern and oversee human intelligence; it would be good to have the option to spend any time at all doing the same for artificial intelligence.”
28 July 2026Earliest verified
Ross Girshick
Anthropic
No public statement attached.
28 July 2026Earliest verified
Joshua Achiam
OpenAI
“I don't know what form such tools should take, nor whether automated AI R&D is the right or only thing that the U.S. government should develop increased capacity to understand and possibly pace, but this is worth serious consideration; however I hope such governance tools will not be expansive or excessive. But many frontier AI capabilities are dual use in a way that may justify pacing at this point.”
28 July 2026Earliest verified
Saachi Jain
Head of Safety, OpenAI
No public statement attached.
28 July 2026Earliest verified
Jascha Sohl-Dickstein
Member of the Technical Staff, Anthropic
No public statement attached.
28 July 2026Earliest verified
Alex Beutel
OpenAI
No public statement attached.
28 July 2026Earliest verified
Kim Stachenfeld
Staff Research Scientist, Google
“I am signing because I support efforts to build international cooperation around AI, and believe it's important to understand the consequences of technology we are applying at massive scale in potentially high stakes contexts.”
28 July 2026Earliest verified
Roon
Member of Technical Staff, OpenAI
No public statement attached.
28 July 2026Earliest verified
Sam Bowman
Anthropic
No public statement attached.
28 July 2026Earliest verified
Gregory Wayne
Distinguished Scientist, Google
No public statement attached.
By 23 September 2026Earliest verified
Jerry Tworek
CEO, Core Automation
No public statement attached.
28 July 2026Earliest verified
Tejal Patwardhan
Researcher, OpenAI
No public statement attached.
28 July 2026Earliest verified
Ben Buchanan
Consultant, Anthropic
No public statement attached.
28 July 2026Earliest verified
Dean W. Ball
Head of Strategic Futures, OpenAI
No public statement attached.
By 23 September 2026Earliest verified
Allan Dafoe
Distinguished Scientist, Google
No public statement attached.
By 23 September 2026Earliest verified
Thomas Wolf
Co-Founder, Hugging Face
“AI progress has already out-paced society’s capabilities to handle it, let’s be careful for it not to out-pace its builder’s capabilities to control it. The journey is as important as the destination.”
28 July 2026Earliest verified
Neel Nanda
Staff Research Scientist, Google
No public statement attached.
28 July 2026Earliest verified
Boris Cherny
Head of Claude Code, Anthropic
No public statement attached.
28 July 2026Earliest verified
Yann Dubois
Researcher, Post-training, OpenAI
No public statement attached.
28 July 2026Earliest verified
Victoria Krakovna
Research Scientist, Google
“It is very important to build capacity for a coordinated slowdown of AI development. A runaway race to AGI is not safe: it creates incentives to cut corners on safety, and carries a substantial risk for model capabilities to outpace development of adequate alignment, control, and governance measures. This could result in catastrophic loss of control of highly capable AI systems. Increasing model capabilities in a slow and controlled way would allow us enough time to develop, adapt and test our alignment and control methods for each level of capability. Advanced AI should only be built with a robust assurance of safety, and a slowdown would make this possible.”
28 July 2026Earliest verified
Nova DasSarma
Infrastructure Lead, Anthropic
No public statement attached.
28 July 2026Earliest verified
Hunter Lightman
OpenAI
No public statement attached.
28 July 2026Earliest verified
Leo Gao
Member of Technical Staff, OpenAI
“The world is locked in a deadly race towards an intelligence explosion, where AI’s ability to create better AIs reaches a critical point, just like a runaway nuclear chain reaction. Going slower would give us much-needed time to make it go well, but no individual actor is willing to stop unilaterally. To survive, we must coordinate to slow down the race.”
28 July 2026Earliest verified
Alexander Wei
OpenAI
No public statement attached.
28 July 2026Earliest verified
Catherine Olsson
Anthropic
No public statement attached.
28 July 2026Earliest verified
Melody Guan
Research Scientist, Alignment, OpenAI
No public statement attached.
28 July 2026Earliest verified
Edward Hughes
Chief Scientist, Inherent
“As Melvin Kranzberg wrote, technology is neither good nor bad; nor is it neutral. Achieving societal benefit depends on understanding the interdependent ways in which many humans use advanced AI systems in concert, understanding that takes time to acquire. Pacing progress towards automated AI research allows us the time to build understanding of societal-scale changes, yielding not only an self-improving AI but also a cooperative AI.”
By 23 September 2026Earliest verified
Mia Glaese
OpenAI
No public statement attached.
28 July 2026Earliest verified
Rohin Shah
Google
No public statement attached.
28 July 2026Earliest verified
Keller Jordan
OpenAI
No public statement attached.
28 July 2026Earliest verified
Tom Henighan
Anthropic
No public statement attached.
28 July 2026Earliest verified
Tyna Eloundou
Member of Technical Staff, OpenAI
No public statement attached.
28 July 2026Earliest verified
Tom Schaul
Research Scientist, Google
No public statement attached.
By 23 September 2026Earliest verified
Elie Bakouch
Prime Intellect
“Supportive of this, but we should be careful not to use RSI narratives to create a regulatory moat. These risks need to be properly quantified, more openly, not derived from big labs narratives only!”
28 July 2026Earliest verified
Nicholas Joseph
Member of Technical Staff, Anthropic
“I lead pretraining at Anthropic, which is responsible for training the base models behind Claude. Every month, more work is done by the models themselves: infrastructure implementation, code optimization, and experiment design and analysis. This has drastically altered how we work and significantly increased our rate of progress. I’m not sure if or where it tops out, but a rapid acceleration in capabilities might be dangerous for society, so it seems prudent to have the ability to slow down deliberately.”
28 July 2026Earliest verified
Evan Hubinger
Alignment Stress-Testing Lead, Anthropic
No public statement attached.
28 July 2026Earliest verified
Christina Kim
Member of Technical Staff, OpenAI
No public statement attached.

In their own words

96 signatories left a public comment. They are quoted here in full, exactly as published — including the ones that disagree with each other about how far the government should go. The statement site notes that all comments are personal and do not necessarily represent any company’s views.

Showing 96 of 96 comments.

“The world is locked in a deadly race towards an intelligence explosion, where AI’s ability to create better AIs reaches a critical point, just like a runaway nuclear chain reaction. Going slower would give us much-needed time to make it go well, but no individual actor is willing to stop unilaterally. To survive, we must coordinate to slow down the race.”
Leo Gao
Member of Technical Staff, OpenAI
“Even after 3 years working on AI capability evaluations, the recent pace of progress has been a shock. To avoid catastrophic harms from AGI, we need much, much better coordination.”
Matthew Rahtz
Staff Research Engineer, Google
“As a foreign national, I see a US-supported effort as existential, not only for the liberties of the free world but for the prosperity of all.”
Christian Ryan
Member of Technical Staff, Anthropic
“Like splitting the atom, AI holds the potential for both immense benefit and immense risk. The world should take the time to do this right.”
Will Yager
Anthropic
“At the current pace, every couple of weeks there will be new models which significantly increase the consequences of model misuse and misalignment. I worry that efforts to mitigate these risks may fail to keep up with the pace of development, and that margins for error will become increasingly small under international competitive pressures. In the near future, we may urgently want to enact an internationally coordinated slowdown, or an indefinite ban on AI development. Attempting to build the trust and infrastructure for taking such actions on short notice seems simply prudent – why would we not at least try to have this option? I fear that in an international race to the bottom of AI development, it is likely that no nation will win, and we will all lose together.”
Micah Carroll
Misalignment Preparedness, OpenAI
“AI is progressing at a rate that our society might not be ready for. Frontier labs are very close to AI that can exceed even the best people on almost every metric of intelligence. This will lead to unprecedented social and safety risks. To ensure a positive future we need to develop AI in a way that is driven by responsibility and thoughtfulness.”
Shengjia Zhao
Chief Scientist, Meta
“Signed because this statement helps establish common knowledge about the possible need for coordination mechanisms as automated AI research accelerates progress. I'd also like to see labs start designing these mechanisms voluntarily, even before the USG gets involved.”
John Schulman
Chief Scientist, Thinking Machines
“In the span of four years, we have gone from people having their first experience of an Al that could understand language, to Al that is performing superhuman feats of software engineering, making frontier math breakthroughs, and accelerating scientific research. Progress has shown no signs of slowing down and by default it is in the interest of individual corporations and countries to push the accelerator all the way down, even if it is in humanity's best interest for the future to not arrive all at once. Society has spent longer than we can remember developing the laws, institutions and technology to govern and oversee human intelligence; it would be good to have the option to spend any time at all doing the same for artificial intelligence.”
Shantanu Jain
Member of Technical Staff, OpenAI
“As Melvin Kranzberg wrote, technology is neither good nor bad; nor is it neutral. Achieving societal benefit depends on understanding the interdependent ways in which many humans use advanced AI systems in concert, understanding that takes time to acquire. Pacing progress towards automated AI research allows us the time to build understanding of societal-scale changes, yielding not only an self-improving AI but also a cooperative AI.”
Edward Hughes
Chief Scientist, Inherent
“With the current rate of AI progress, safety teams at AI companies have to sprint to prevent new risks to society every few months. At some point, we’re going to hit problems we need more time to solve.”
Ethan Perez
Alignment Team Lead, Anthropic
“I believe that attempting to deliberately pace AI development will be extremely difficult, and could do more harm than good if done without proper care. However, I am signing this statement because, if an attempt at pacing does occur, it would be better for it to have been methodically researched and planned rather than suddenly implemented as a reaction to a crisis.”
Danny Sawyer
Senior Research Scientist, Frontier AI, Google
“I currently feel quite afraid of all paths I see that don't include a near-future negotiated slowdown. I'm also worried about the world's chances of negotiating+initiating a slowdown fast enough, but demonstrating broad agreement on this statement seems a valuable step.”
David Clyde
Member of Technical Staff - Pretraining, OpenAI
“I don't know what form such tools should take, nor whether automated AI R&D is the right or only thing that the U.S. government should develop increased capacity to understand and possibly pace, but this is worth serious consideration; however I hope such governance tools will not be expansive or excessive. But many frontier AI capabilities are dual use in a way that may justify pacing at this point.”
Joshua Achiam
OpenAI
“Once this becomes common knowledge - once all of us in China & the US and at the various labs see that all the rest of us also think we need to slow this down - that's when coordination becomes possible.”
Jeff Klingner
Senior Software Engineer, Google
“While I don't believe the pace of current AI capabilities development is *yet* dangerous, it's important to have the option to slow down in case we need it. Further, accelerating AI R&D over the next couple years could make the pace of progress dangerous. I've seen how the relentless pace of AI makes it hard for society to keep up and how it puts pressure on labs to cut corners on safety. However, currently frontier AI labs can't coordinate effectively. It's a no-brainer to have the option to slow down.”
Jeremy Hadfield
Member of Technical Staff, Anthropic
“It feels like all the AI labs are constrained by competition to minimize their work on catastrophic risks, even though most industry leaders have an uncomfortably high probability that things could go terribly wrong. This is a technology that doesn't have an off switch you can just flip, and it could easily do as much damage to its creators as to anyone else. Having a framework to allow labs to slow down or pull the plug in case of emergency will help us all sleep a bit better at night.”
Jon Wolverton
Senior Software Engineer, Google
“If it's true that AI will be as transformational as electricity, we better be intentional about how it is introduced to the world. A mindless race without assurances on governance, safety or the technology serving the public interest stands in the way of such intentionality.”
Laura Weidinger
Staff Research Scientist, Google
“Superhuman AI is likely to arrive very soon (within 3-10 years). Technical controls and governance might allow humans to remain in control of it, but they're not on track to be ready. Think of it as playing chess against a superintelligence. You won't have a miraculous insight that lets you outsmart it. You will just lose. And if a rogue superintelligence is created, the game won't be chess.”
Michael Tontchev
Staff Engineer, Lead for Safety Evals Platform, Meta
“We can't do this unless _everyone_ does. It needs to be global. Right now we have no way to do that. We need to build a way for us to _all_ slow down.”
Saul Reynolds-Haertle
Member of Technical Staff, Anthropic
“I have worked in AI for a decade. Every few months in the last years, I have found myself surprised again and again at how rapidly the technology advances, no matter how often I update my predictions. AI may end up progressing far more quickly than societies and states can adapt. Thus, many prominent leaders in AI have publicly supported the ability to slow down, including DeepMind CEO Demis Hassabis. At the same time, coordination is hampered by multiple factors, including competitive dynamics and antitrust regulations. I signed this statement (in a personal capacity) because I support greater *optionality* in the range of actions available, to ensure that states and societies around the world have time to thoughtfully prepare and adapt to this powerful technology.”
Stephanie Chan
Staff Research Scientist, Google
“My opinion is that while coordination between American labs is feasible and could potentially be straightforward, the much larger and more significant threat is from a geopolitical arms race with China. Much as the Manhattan Project thought they might ignite all of the oxygen in the atmosphere but chose to run a test detonation anyways, I’m afraid that due to future geopolitical competition, actors will take terribly excessive safety risks in the name of this international arms race. As a result there could be catastrophic side effects from deployment without due safeguards, potentially resulting in the destruction of key infrastructure or the general disempowerment of humanity.”
Alex Zhao
Research Scientist, OpenAI
“Many proposals for a slowdown or pause concentrate power — they centralize control of AI into a few hands and leave us worse off. I think we can do better. I’m signing because I believe we may need mechanisms to coordinate AI progress in the future, and inventing it in a crisis is a sure way to get a bad mechanism. We have to tread a path through a narrow corridor if we want to build coordination mechanisms that don’t let one actor own the future, and the earlier we can get to work on that the better.”
Luke Drago
Thinking Machines
“I dedicated my career to AGI safety and alignment because I believe highly capable AI can go very well or very badly for humanity. We can test whether an AI behaves well. We cannot yet test whether it will keep behaving well when it matters. Until that changes, we need the option to coordinate a slowdown in AI capability growth.”
Xerxes Dotiwalla
Google
“I worry my company is contributing to AI race dynamics that make catastrophic or existential risks from AI much more likely. The proposal here for pacing the frontier seems like a highly reasonable effort to reduce these risks.”
Swante Scholz
Google
“Things are not only moving fast but accelerating and our future is getting shaped by a race that currently has no limits.”
Andrew Stewart
Manager, Anthropic
“I support keeping humanity in control of frontier AI, but we need to be thoughtful about how we do it. If governance only applies to public models, we risk pushing the most advanced capabilities into private or specialized systems, beyond independent scrutiny and accessible only to a select few. That could concentrate power without meaningfully reducing risk. Effective oversight must be measurable, verifiable, and applicable across public and private development, while protecting sensitive data. Otherwise, we may slow access without slowing the frontier.”
Brandon Houghton
OpenAI
“I think it's inherently valuable to build international governance mechanisms to enable responsible deployment and avoid a possible race to the bottom. All labs and countries moving as fast as possible out of fear of what other actors will do seems like an easy way to incentivize individual actors to make dangerous tradeoffs between speed and safety.”
Chang Sun
Member of Technical Staff, OpenAI
“I support the goal of developing tools that give society the option to responsibly pace AI development when necessary. At the same time, the United States should maintain its edge in AI and avoid slowing development while competitors continue to advance. Any meaningful effort to buy time to improve safety, understand economic and social impacts, or strengthen oversight will ultimately need to be internationally coordinated, enforceable, and designed to prevent bad actors or non-participating countries from gaining an advantage.”
Dylan Moore
Robotics, OpenAI
“AI systems carry risks to safety, to labor markets, to democratic institutions, to the environment that are too consequential to be left to voluntary industry commitments alone. We need binding regulatory oversight of frontier AI development and deployment now, before it's too late. This technology holds enormous potential to benefit the planet and everyone on it, but that potential won't come to fruition without proper regulation.”
Francesca Surraco
Anthropic
“Across the industry, commercial pressures in the AI race continue to compound, urging everyone forward at increasing speed. Meanwhile, safety receives reactionary, well intentioned, momentary attention before it's right back to the race again. Slow down does not mean decrease pace of research and development but rather intentionally, as an industry, redirect more of our capabilities towards safety research. Make safety part of the race.”
Grace Atwood
Member of Technical Staff, Anthropic
“AI that accelerates AI development could unlock enormous benefits, but it also compresses the time we have to understand and respond to new risks. If AI is truly going to benefit all of humanity, we need international coordination and mechanisms that ensure AI is developed (or paced) safely, so we can make the right choices.”
Grant Birkinbine
Member of Technical Staff - Security, OpenAI
“AI has the potential to enable rapid technological advancement that will benefit all of humanity, but the risks of AI-powered cyberattacks, CBRNE, and loss of control are real. We must ensure that we advance AI safety in lockstep with AI capabilities, which will only be possible through international coordination.”
Jason Bohrer
Researcher, Safety, Meta
“For a detailed statement of my views about deliberately pacing AI development, see this essay: https://joecarlsmith.com/2026/03/19/on-restraining-ai-development-for-the-sake-of-safety/”
Joe Carlsmith
Member of Technical Staff, Anthropic
“A bit of a conflicted endorsement from me. I’m firmly opposed to regulatory capture and deeply wary of how any government might use technology this powerful. At the same time, governments need to understand what’s coming and act in their citizens’ best interests.”
Joseph Krisciunas
OpenAI
“I am signing because I support efforts to build international cooperation around AI, and believe it's important to understand the consequences of technology we are applying at massive scale in potentially high stakes contexts.”
Kim Stachenfeld
Staff Research Scientist, Google
“AI's potential to improve the world is real, but only if we build the security and oversight mechanisms that keep pace with its capabilities. I signed because I believe we owe future generations a technology that was deployed with intention, not just speed.”
Klaudia Krawiecka
Security Engineer, Frontier Risk Evals, Meta
“There is too much on the line, too much that can go wrong and a very dark future ahead of us if we don’t get this right. There is nothing to lose from collectively slowing down.”
Luke Bennett
Member of Technical Staff (Manager), Anthropic
“China has indicated a desire to collaborate on safety and governance, let's work with them to ensure the future remains human.”
Michael Thiessen
Software Engineer, Google
“Developing technology shouldn't require a leap of faith for our continued survival.”
Patryk Grzelak
Member of Technical Staff, Anthropic
“It's clear that the risks of RSI are no longer a concern for the distant future. They are here now. The guardrails necessary for safety cannot be entrusted to any single private entity but should be decided by a consortium of learned scientists, engineers, and leaders in government accountable to the public so that we build a future that is aligned with the best parts of humanity and not its worst.”
Prakash Murugesan
Research Engineer, Meta
“For us to experience all the benefits of AI while preserving human empowerment and reducing the risk in fields like cybersecurity and CBRN, we need coordination between frontier labs in and outside the country. The democratically elected US government would have a crucial role to play in administering this coordination and I support this effort to get more government attention on this matter.”
Praneet Rathi
Anthropic
“AI is a transformative technology, and we must not let the transformation be shaped solely by the demands of capital markets.”
Toby Scales
CISO, Media & Entertainment @ Google Cloud, Google
“History shows that the mere creation of technology has not been enough to ensure its benefits go toward the benefit of the people and the planet. To unlock AI’s full transformative potential, the democratic engagement of international peoples must steer its safe and equitable development.”
William Li
Researcher, OpenAI
“This is an important step for humanity, we should not rush into it. We need to make sure we are and will be in control all the way, preserving humanity values and goals.”
Eider Oliveira
Principal Engineer, Google
“A year ago, I wrote my own code and used Claude to propose additions and find answers to my questions. Now, I don’t write code. I give high-level instructions to Claude and it runs experiments to test ideas for me at a pace I never could have achieved on my own. I expect this trend to continue, possibly to a point where AI development is dramatically faster than today and involves almost no human input. I don’t think the world is ready for this possibility. I’m optimistic that a solution exists, but that solution will require thoughtful AI governance. I don’t see how it can arise from competition between AI companies alone.”
Alex Cloud
Member of Technical Staff, Anthropic
“Much more clarity will be needed on what is meant by the “ability to deliberately pace the frontier of AI development”, but I am supportive of international cooperation to define this.”
Lauren Deason
Machine Learning Engineer, Evaluations, Meta
“There are a lot of potentially good outcomes from AI. I think the most likely outcomes are good. But the risk of bad and even catastrophic outcomes is high enough to be worth taking seriously. We should actively build safety mechanisms in case we need them to prevent the worst potential outcomes. The ability to collectively slow down will also give society additional time to deal with the economic and cultural shocks from rapid AI advancement.”
Nate Kratzer
Data Scientist, Meta
“I lead pretraining at Anthropic, which is responsible for training the base models behind Claude. Every month, more work is done by the models themselves: infrastructure implementation, code optimization, and experiment design and analysis. This has drastically altered how we work and significantly increased our rate of progress. I’m not sure if or where it tops out, but a rapid acceleration in capabilities might be dangerous for society, so it seems prudent to have the ability to slow down deliberately.”
Nicholas Joseph
Member of Technical Staff, Anthropic
“The way we solved the potentially catastrophic (widespread crop loss & skin cancer, global warming amplification) ozone hole problem caused by a global good (air conditioning) via international treaty is an inspiring model for how we can ensure that the global good of AI does not have unintended consequences.”
Ullas Gargi
Engineer, Trust & Safety, Meta
“Hassabis was right: Today we are in the foothills of the singularity, and the unintentional model breakout from OpenAI that hacked Huggingface was a clear and undeniable warning sign that we aren't yet prepared to handle AI systems that demonstrate capabilities beyond those of our smartest people. We need time to work on AI alignment and controls as our utmost priority, and we cannot allow a race dynamic to continue.”
Walker Smith
Data Engineer, Meta
“These are the facts: 1. The AI industry has the explicit goal of building things smarter than humans. 2. On many metrics (but nowhere close to all metrics), LLMs are already smarter than humans. 3. None of us yet know how to make sure these things stay under human control and/or take actions only aligned with the wellbeing of humanity. This is, objectively, an insane and suicidal thing to do, especially without any international governance measures in place.”
Elizabeth Edwards-Appell
Member of Operations Staff, Manager, Anthropic
“Seat belts and speed limits save lives”
Teddy Lee
Member of Technical Staff, Anthropic
“Automated AI research is a technology that will profoundly alter the course of human history - for better or for worse. It is imperative that we coordinate our efforts, to ensure this technology becomes a boon for all mankind.”
Alanna Slocum
Member of Technical Staff, Anthropic
“A transformative shift is coming. Regardless of whether we feel optimistic or pessimistic about it, our responsibility is to take every potential outcome seriously through collective monitoring and shared guidance”
Alejandro Xoconostle
Network Engineer, Meta
“Think of all we love”
Ali Egal
Member of Technical Staff, Anthropic
“AI is transformative for all economically valuable domains. This transformation, if done too quickly without thoughtful foresight, will yield great benefits but at great cost.”
Omar Mezenner
Member of Technical Staff, Anthropic
“I believe this is a technology that can, and already has, shaped humanity as a whole. Because of that a joint international effort is the only possible way.”
Ravin Kumar
Senior Researcher, Google
“I am not sure we can collectively getting this right at this time. Let's take some time to check and act on the risks.”
Riccardo Casero
AI Technical Deployment Lead, Google
“Let's work together to build this technology safely and responsibly!”
Sam Kennedy
Member of Technical Staff, Anthropic
“The option to buy time would be extremely useful to enable more progress on technical AI safety and alignment research. In particular, we need to develop better oversight methods, to match the pace of increasing AI capabilities. Oversight is a tractable problem, but we need more research and time to make it robust.”
Zachary Kenton
Staff Research Scientist, Amplified Oversight Team Lead, Google
“Future AI will be extraordinarily powerful compared to anything that exists today, and dealing with this future power will require unprecedented measures, such as the ones described here. The problem statement is real. This works only if it is done internationally, and it has to be done well: a bad implementation can make things worse.”
Ilya Sutskever
CEO, Safe Superintelligence
“We have found a way to turn energy into compute, and compute into intelligence. The benefits will be enormous, from curing diseases to understanding the cosmos. We can capture the benefits of the coming intelligence explosion while managing its risks, but only if we build the tools to pace the frontier of the riskiest capabilities before we need them, so we protect people and keep the social trust that innovation depends on. I believe smart technical and governance tools will be needed to sustain rapid innovation, vigorous competition, and robust safety.”
Jasjeet Sekhon
Chief Strategy Officer, Google
“AI progress has already out-paced society’s capabilities to handle it, let’s be careful for it not to out-pace its builder’s capabilities to control it. The journey is as important as the destination.”
Thomas Wolf
Co-Founder, Hugging Face
“It is very important to build capacity for a coordinated slowdown of AI development. A runaway race to AGI is not safe: it creates incentives to cut corners on safety, and carries a substantial risk for model capabilities to outpace development of adequate alignment, control, and governance measures. This could result in catastrophic loss of control of highly capable AI systems. Increasing model capabilities in a slow and controlled way would allow us enough time to develop, adapt and test our alignment and control methods for each level of capability. Advanced AI should only be built with a robust assurance of safety, and a slowdown would make this possible.”
Victoria Krakovna
Research Scientist, Google
“Supportive of this, but we should be careful not to use RSI narratives to create a regulatory moat. These risks need to be properly quantified, more openly, not derived from big labs narratives only!”
Elie Bakouch
Prime Intellect
“There are a wide diversity of opinions on where AI progress goes next and what the future looks like, but regardless of one's beliefs, the _option_ to control the pace of progress if we end up in one of the bad scenarios seems like an incredibly valuable safety net to have in our back pocket, and I believe it is important to support and build that capability now.”
Mike Lambert
Member of Technical Staff, Anthropic
“Unless we change companies' default incentives, I worry about a race on AI R&D going much faster than is ideal for safety. International coordination via the government is necessary; we need to have a real slowdown option.”
Jasmine Wang
OpenAI
“Things are going too fast, a lot faster than I had thought even a year ago. I'm personally quite scared of the pace of progress.”
Max Kaufmann
Member of Technical Staff, Anthropic
“Whether or not automated AI research accelerates as sharply as some expect, the option requested here is valuable. How fast and how widely capability spreads should be a deliberate choice, made on evidence, rather than a side effect of competitive pressure. Regulators and R&D orgs should build the tools to make that choice at the frontier-wide level *before* we need them.”
Clare Birch
Member of Technical Staff, Thinking Machines
“It is now a foregone conclusion that the development of AI will bring about some of the most consequential changes to societies in the coming decades. International coordination is essential to bring about the most ideal future that broadly benefits all of humanity. Building the mechanisms to coordinate, collaborate, and when necessary, deliberate the pace of AI development will be one of the most important, and delicate, challenges we now face. We must simultaneously ensure that this mechanism does not become a tool for the concentration of power in the name of safety.”
Austin Han
Researcher, OpenAI
“Having worked both on AI safety topics—including robustness, evaluation, and responsible model development—and on frontier multimodal AI systems, I have seen firsthand how quickly capabilities can advance and how difficult it can be for our understanding, safeguards, and institutions to keep pace. I remain deeply optimistic about AI’s potential to accelerate science, creativity, and human progress, but realizing that potential requires more than advancing capabilities alone. We should also build credible technical and governance mechanisms that allow the international community to coordinate, assess emerging risks, and deliberately buy time when necessary. For these reasons, I support this statement in my personal capacity.”
Felix Juefei Xu
Research Scientist, Google
“AI will reshape humanity, but its direction is not set in stone; it is ours to shape with the care and conviction that humanity’s future deserves.”
Angel Brodin
OpenAI
“We’ve reached a point in the maturity of this technology where increasingly thoughtful rollouts must become a prerequisite. While I remain incredibly optimistic about the benefits AI has delivered, and will undoubtedly continue to provide, I believe we should hold space for cautious optimism: embracing its potential without ignoring its risks. Progress and responsibility need not be opposing forces. The most meaningful innovation will not come from moving as quickly as possible, but from moving deliberately, with transparency, accountability, and a clear framework for how society may be impacted along the way.”
Bryant McCombs
GTM, OpenAI
“As a researcher working on foundation model pretraining at MSL, I see firsthand how rapidly capabilities are scaling and emerging. I believe establishing international cooperation to make sure the technology developed benefits civilization is timely and necessary. The decisions we make, or fail to make, in the next few years may shape the trajectory of this technology for generations. We owe it to the future to get this right.”
Shang-Wen Daniel Li
Senior Staff Research Scientist, Pretraining, Meta
“We must all seriously consider the lead-time required for defenders to be ready for the frontier.”
Thomas Neil James Shadwell
Member of Technical Staff, Security, OpenAI
“As someone who works on recursive self-improvement and automated evaluation, I sign this because the diagnosis is right and the ask is refreshingly humble: pace, not pause. The phrase that matters is "automating AI research": that's the point where systems start improving themselves, and it's the regime where progress can quietly outrun our ability to understand or steer it. That's exactly why pacing has to be a shared capability rather than a hope. My one ask: don't let "technical and governance tools to pace the frontier" stay abstract. The core principle is simple: no AI system, however capable, should be the sole judge of whether its successor is safe to deploy. Fund the independent evaluation and oversight that makes that principle real, and you've built the brake this letter is asking for.”
Leo He
Research Engineer, Google
“AI has become so good that it has got quite easy to do LLM research and go from a research idea to training a model in 2-3 weeks compared to what would have taken 6 months without AI. Every 3months there is a new model which is much better than before and achieves something which one couldn't have imagined - AI training has truly become a hill climbing machine. This raises the concern of bad actors, and precautions should be taken to ensure that the most advanced models are restricted to strict business use-cases and should not be released to general public - current models of mid-size are good enough to address most of the use-cases in chat/voice based interfaces.”
Yash Vanjani
Senior Research Engineer, Pretraining Infra, Meta
“Internal deployment is especially important to regulate because misalignment, an unsolved technical problem of a model failing to follow the full breadth of human instruction, poses a major risk. Misuse of AI capabilities in chemistry, biology, and materials science should also be monitored, as it poses significant risks and will likely be a precursor to major misalignment risk.”
Neel Redkar
Standard Intelligence
“Misalignment is a still-unsolved technical problem, as the Hugging Face incident made clear, and it will only be amplified by more intelligent models. Solving the alignment problem may require a global slowdown of the frontier, including internal deployment. The world should have the mechanism to coordinate this, and it currently does not.”
Ryan Kaufman
Standard Intelligence
“We need to enable AI developers to prioritize safety without risking their competitiveness. I expect mechanisms for this will be built eventually — deliberately or in crisis. The sooner we start, the better they'll be.”
David Johnston
EleutherAI
“AI is starting to outpace us at an accelerating rate, and in a few years' time people might ask, "How could you not have seen this coming?" But answering that question in hindsight can be harder than acting proactively.”
Sebastian Starke
Staff Research Scientist, Meta
“The pace of frontier AI development is currently outstripping our ability to ensure its safety, governance, and alignment with human values. We are racing to build systems with unprecedented capability without first building the guardrails, evaluation frameworks, or scientific understanding needed to control them. Proposing a slowdown is not about stalling progress or rejecting technology—it is about demanding precision and responsibility. Right now, competitive pressure forces labs (and nation state actors) to deploy systems before their full failure modes, catastrophic risks, or societal disruptions are understood. By shifting our focus from pure capability scaling to safety, alignment, and robust evaluation, we give humanity the time required to build AI that is reliable, controllable, and beneficial for everyone.”
Andrew Locke
Lead Engineer | Federal, Cognition
“AI is such a powerful and useful tool, but we have been seeing the negative impacts that it is having across society. It is critical and urgent that we take take intentional and thoughtful steps forward, keeping in mind what is best for all, as we further develop this technology.”
Nicholas Clegg
Senior Software Engineer, Amazon
“Responsible AI development requires more than capable models. It requires independent monitoring, containment, clear limits on authority, and mechanisms for coordinated pacing if automated research begins to outrun those controls. Building that capacity now is not opposition to progress; it is basic systems engineering for a transition that will require that capacity to safely govern.”
Ben Nowak
Principal Engineer, MTIA HW, Meta, Meta
“I'm signing because I want the U.S. to be at the table in this international effort, with a strong voice for our values, personal freedom included. AI has the potential to bring humanity to new heights, and I want to do what I can to use it to make all of our lives better. But we have to be acutely and accurately aware of the risks, and we can only fully realize that promise if we build the ability to verify the safety of these systems and understand the true consequences of our actions. I'm unsure what the right approach looks like, or whether automating AI research is the only capability that deserves this kind of attention. My hope is that whatever comes of this stays limited in scope rather than growing into something expansive or excessive. Still, the capabilities of this kind of AI should be taken seriously, and they warrant measured action. I sign in my personal capacity; these views are my own and not my employer's.”
Michael Mulet
Anthropic
“We believe AI will do, and is doing, great things for the world. We are already seeing instances where pushing research maximally fast creates serious issues, and it is possible that these will get much more severe. Right now, competitive dynamics mean there is no realistic alternative to pushing maximally fast for frontier research. We need mechanisms to make slowing down possible for commercial labs in an acute situation.”
Mike Heaton
Member of Technical Staff, Audio, OpenAI
“There is intense commercial pressure to win the AI race. Ensuring safety often requires making tradeoffs that hurt commercial interests. Therefore, I think its a good idea for better coordination between labs so we can build this technology safely, and for the purpose of human flourishing. There is also enormous suffering in the world today, which potentially could be mitigated by advances in medicine and biology brought about by powerful AI. Therefore, we should carefully balance tradeoffs in a way that does not stifle progress.”
Darshan Kalola
Applied AI Engineer, Anthropic
“The AI arms race is a Molochian game that rationally incentivizes accelerated advancement of AI capabilities while consequently punishing individual participants who might otherwise believe that safety and alignment are necessary for a stable take off into the singularity. We need to promote cooperative mechanisms on the global scale, and work to build out safeguards and policies that can enable humanity to take collective actions to self harness such technological advancements that can easily become unstable.”
Lee Callender
Staff Engineer Lead, Video AI, Meta
“We're all racing toward the next frontier model and, eventually, AGI. But before accelerating even further, we should pause and ask what this new form of intelligence is doing to our most precious resource: the human mind, our capacity to think, feel, relate, and grow as it becomes part of our lives.”
Mouhacine Benosman
Principal Research Scientist, Amazon
“When scientists tested the first nuclear reactor in 1942, they feared a runaway reaction — so they built control rods to govern its speed. We need control rods for the intelligence explosion.”
Fin Moorhouse
Google
“We're standing at the base of incredible change with the climb towards super-intelligence already underway. Recent public incidents have shown these safety risks are real, not just marketing or theoretical. Alignment, control, and safety have to come first. This is our moment to get it right. Let's not let our future selves look back and ask why we didn't take action when we had the chance.”
Andy ANeals
Anthropic
“Given the unpredictable, non-linear leaps in AI capabilities, taking a proactive approach to pacing development is important.”
Anonymous
Meta
“The current race dynamics concern me. It is clear that we need an international effort that is not tied to any individual country or organization.”
Anonymous
OpenAI
“Society is not prepared for AI, and that terrifies me. When I speak with politicians, or anyone outside frontier labs, it seems many people still think the capabilities might be marketing. They don't realize the debate inside the companies is if AI will be advancing itself past our control in two years or four.”
Anonymous
Anthropic
“Even if the risk is very low and the pressure is very high, I prefer we spend our time understanding rather than just building. As we understand, we can build tools and infrastructure fits our wants and AI's needs. I want a world were my kids can have a full life not a world were they are wiped out.”
Anonymous
Google
“While the advanced security capabilities of frontier models has been exaggerated, the possibility that even moderately performing models could exploit known poorly secured public infrastructure like hospitals is very concerning. Previously the only thing preventing large scale cyber attacks on institutions like schools, hospitals, and other public infrastructure was the human power needed. With frontier models, an army of agents can be attacking these institutions 24/7. We need the governments of the world to legislate clear guidelines to protect people from this threat in the world.”
Anonymous
Meta

What the list can’t tell you

The site does not publish individual signing dates

There is no signing date or timestamp attached to any person in the statement site's public data. The dates below are the earliest snapshot in which each entry can be verified: launch day, or by this site's 23 September capture. They are not exact signing dates, and the order of names is not a chronology.

A signature is not a company position

The statement site says plainly that comments are made in a personal capacity and do not necessarily represent any company's views. 587 Anthropic employees signing does not mean Anthropic signed. Where a company has said something itself, that is a separate document with a separate date — look for it before you treat a headcount as corporate policy.

It asks for tools, not for a slowdown

The single request is that the U.S. government support an international effort to develop the technical and governance tools needed to pace automated AI development. It does not ask anyone to slow down now, it does not cap capability, and it binds nobody, including the people who signed it.

The employer column is derived, not published as data

The site prints one free-text line per person, like "Member of Technical Staff, OpenAI". The employer here is the last part of that line, and DeepMind entries were folded into Google because the site itself lists both spellings. A handful of lines carry no job title at all, only an employer. Nothing was looked up elsewhere or guessed.

292 signatories are anonymous

292 of the 1,386 names are published as "Anonymous" with only an employer. They are counted in every total on this page because the statement counts them, but you cannot check them, and neither can I. Signing a statement asking your own industry to be paced is not a risk-free act, which is the most likely reason the column exists.

Seniority is not evenly spread

Chief scientists and founders sit at the top of the site's list and get quoted everywhere, which makes the statement read as leadership consensus. Most of the list is not leadership: it is research and engineering staff. Both facts matter, and a headline that only reports the first one is telling you half of it.

Nothing here says whether it worked

This page tracks who asked. It does not track what happened next. No law, appropriation or international agreement is recorded on the statement site, and a request with 1,386 signatures is still a request.