We aim to catalyze collective action to ensure humanity survives this decade. See 100+ ways individuals, organizations, and nations can help.
San Francisco, California
Joined December 2024
- Tweets6.1K
- Following254
- Followers895
- Likes7.5K
The Center for Existential Safety believes international security is best achieved through shared prosperity.
Our plan is to catalyze an immediate, indefinite, and international pause on the development of superintelligent AI. There is a win-win way forward that protects and benefits all nations at the same time: TrumpXiDeal.ai.
Then we should work towards stronger international governance of AI, for the good of all life: iaiga.org.
It will be an extremely harrowing few months or years, regardless of what we do or don't do now. But our odds improve the more people act. See 100+ concrete actions individuals, organizations, and nations can take to help: existentialsafety.org.
Join the movement. Do something constructive every day: actionforsafety.org. We should fight hard for eutopia and against dystopia. Both paths are possible right now.
Today we should all act to preserve our civilization once again.
See 100+ ways how individuals, organizations, and nations can help here: existentialsafety.org.
Let's fight peacefully, passionately, and persistently for our collective safety.
Today in 1962, JFK told the public about nuclear missiles the USSR had secretly placed 90 miles from US soil.
Now is an important time to remember the Cuban Missile Crisis because, until now, it was probably the closest the world ever got to ending.
But now there is a new and worse crisis. Most experts agree that there are double-digit odds of superintelligent AI killing every single human. There is no plan in place to prevent this from happening, and no one knows how much time is left.
The AI crisis is harder to explain than a nuclear missile aimed at your home. But people are starting to pay attention, and asking: what are we going to do about this?
So it's probably a good time to look back and remember how humanity has faced other global emergencies.
Frontier AI staff please consider joining our collaborative safety call on October 27th: luma.com/h1sm6kw5.
You hold more power now than almost anyone in human history. With that comes great responsibility and a unique opportunity to do an enormous amount of good.
We'll politely and privately discuss many options you might consider taking during the call.
Please share this post with any frontier AI staff you know.
The current cybersecurity crisis--and plausibly imminent biosecurity crisis--is why some of us established and moved into an off-grid existential safety community on a tropical island.
If you're dedicated to existential safety, come live and work alongside us. Just reach out and we'll discuss the details. Space is unfortunately very limited.
For those who want to set up their own space, see survivalsanctuaries.com. The team will advise anyone who has signed or is acting in the spirit of the Existential Safety Pledge (existentialsafetypledge.org).
We should all help prevent the existential risks we face from superintelligent AI. For most of us, there is no better use of our time. Don't delay. Act now. We list 100+ ways how the public and policymakers can help here: existentialsafety.org.
Start with signing and sharing the Trump-Xi AI Deal on every platform you have: TrumpXiDeal.ai.
X:
xitter.cf/aisafetyaction/status/…
Facebook:
facebook.com/share/p/1DAy7jv…
LinkedIn:
linkedin.com/posts/the-deal-…
Then keep going. The more that join, the better. There is power in numbers.
President Trump and President Xi could make a deal that ends the AI race before AI ends us.
It would be the most win-win deal of all time. A Nobel Prize-worthy achievement.
Sign and share to show your support: TrumpXiDeal.ai.
The near-term catastrophic risks from AI are abundantly clear, if you choose to pay attention to them.
This is an important paper highlighting some of those risks and how they might affect you.
Do not delay on preparing for both the worst case and best case scenarios we collectively face. We list 100+ ways how to both reduce and prepare for the risks here: existentialsafety.org.
🚨New paper (possibly the most important one I have written yet).
We think that as early as 2027 or 2028, there might be a major AI-enabled terrorism event. We conducted a pre-mortem, made 9 falsifiable predictions, and wrote 2 concrete scenarios.
Led by @KamacheeMax.
About time.
I don't usually agree with Republican officials. But yesterday, a group of eight state Attorneys General, including seven Republicans, sued OpenAI for endangering the American people through its reckless development of ChatGPT.
Here is just some of what they said:
The Arkansas Attorney General, Tim Griffin, argued that as safety concerns mounted, “OpenAI did not slow down. It held its foot on the gas, releasing increasingly powerful models without adequate safeguards or testing as growth and profits repeatedly took priority over safety.”
The Montana Attorney General, Austin Knudsen, argued that OpenAI’s “revenues have come at great cost to the public, which has suffered irreparable harm—including lost loved ones, disturbing cyberattacks, and now a plausible threat of civilizational collapse.”
The Iowa Attorney General, Brenna Bird, demanded a PAUSE on AI testing “until adequate safeguards are implemented to ensure that the tests do not pose a danger to the public.”
They're right.
If Big Tech Oligarchs are building a product that is threatening the safety of our children and all of humanity, they must be held accountable – no matter how rich and powerful they might be.
The UN has released a new report on the advancement of AI, stating that artificial general intelligence (AGI):
• Might be developed soon: “Industry leaders and experts anticipate AGI could emerge within this decade.”
• If handled poorly could result in irreversible consequences for humanity: “Once AGI is achieved, its impact may be irreversible. With many frontier forms of AI already showing deceptive and self-preservation behavior, and the push towards more autonomous, interacting, self-improving AIs integrated with infrastructures, the impacts and trajectory of AGI can plausibly end up being uncontrollable.”
• Could lead to the creation of weapons of mass destruction: “AGI could enable some states and malicious nonstate actors to build chemical, biological, radiological, and nuclear weapons. Moreover, large, AGI-controlled swarms of lethal autonomous weapons could themselves constitute a new category of WMDs.”
• Could get out of humanity’s control, posing existential risk: “AGI might also seek power as a means to ensure it can execute whatever objectives it determines, regardless of human intervention. National governments, leading experts, and the companies developing AGI have all stated that these trends could lead to scenarios in which AGI systems seek to overpower humans. These are not far-fetched science fiction hypotheticals about the distant future—many leading experts consider that these risks could all materialize within this decade, and their precursors are already occurring.”
In response, the UN proposes:
• The creation of a Global AGI Observatory "to track progress in AGI-relevant research and development and provide early warnings on AI security to Member States.”
• An International System of Best Practices for AI development: “Given that AGI might well be developed within this decade, it is both scientifically and ethically imperative that we build robust governance structures to prepare both for the extraordinary benefits and extraordinary risks it could entail.”
• A new UN Framework Convention on AGI “to establish shared objectives and flexible protocols to manage AGI risks and ensure equitable global benefit distribution.”
Press release: uncpga.world/wp-content/uplo…
This would greatly help with our collective existential safety and we support it. But given the existing and likely future cases of loss of control of frontier AI, we are calling for an immediate, indefinite, and international pause on development of superintelligent AI, including an immediate ban on all recursive self-improvement (RSI) methods.
We propose two actions the public can take now to help make this happen:
• Sign the Trump-Xi AI Deal. A win-win deal between the United States and China would be a Nobel Peace Prize-winning move that ends the riskiest parts of the AI race: TrumpXiDeal.ai.
• Sign the International AI Governance Alliance petition. We need to create an international body to govern AI as quickly as possible: iaiga.org.
Please take action today. We list 100+ ways how individuals, organizations, and nations can help move us towards a safer future for all: existentialsafety.org.
It's fantastic to see solid research back the commonsense idea of pausing dangerous AI.
Humanity does hard things all the time. This needs to one of those times.
#PauseAI
We signed and suggest you consider doing so, as well.
The AI Pledge for Humanity has now surpassed 1,000 signers! Join the growing chorus of people supporting an AI dividend and also those showing genuine support for UBI by commitments to fund UBI orgs and projects.
Add your name at aipledgeforhumanity.org and let others know you signed it and why.
ALT Graphic for the AI Pledge for Humanity, dated October 2026, with large text announcing “1,000+ signers. Across 50+ countries.” Below: “AI was built from all of us. The AI dividend belongs to all of us.” Featured signers include Vernon L. Smith, Nobel Laureate in Economics; Andrew Yang, former presidential candidate; Dustin Moskovitz, co-founder of Facebook and Asana; and Guy Standing, co-founder of BIEN. Footer reads: “Add your name: aipledgeforhumanity.org” and “Income To Support All Foundation.”
To frontier AI staff, you have a moral obligation to fix the problems you have helped create. Here are several concrete actions you could consider taking: xitter.cf/aisafetyaction/status/….
We'd like to see thousands of you join the ever growing community of former frontier AI staff who have left and courageously spoken out about what you witnessed: ethicalaidepartures.fyi.
You can truly make a profound difference for humanity.
Act now. The world is waiting. And watching.
We have a choice in front of us. In my latest op-ed for Transformer, I invite employees of frontier AI companies to ask themselves if their work is truly leading us toward a safer future. Reflecting on these questions is hard, but we no longer have the luxury of time. We must act now.
Read my full op-ed: [transformernews.ai/p/yoshua-…]
Now is a uniquely powerful time to leave and whistleblow. A tipping point could be reached very quickly, leading to substantial changes at the companies themselves and in the geopolitical arena.
1. It would signal to your counterparts around the world that the risks are indeed real and their leadership needs to also act now. US and Chinese political leadership in particular needs to be convinced.
2. Your former company might be forced to become genuinely safer lest they lose even more staff and investors.
3. It would encourage the public to take AI safety much more seriously, and vote accordingly.
The AI race, which many of us consider tantamount to suicide, is not inevitable. For example: xitter.cf/aisafetyaction/status/…. Humanity can find a better way, together. With extinction or eutopian scenarios on the table, an enormous amount is now possible.
On that note, here are 100+ concrete ways anyone inside or outside the AI companies can help with existential safety: existentialsafety.org. Millions have already acted. You can be busy, if you choose. And be in good company.
Before you make up your mind, consider doing these three things:
1. Read this excellent article on why people should generally not work at frontier AI companies: intelligence.org/2025/06/11/….
2. Deeply examine what your companies have been up to: openaifiles.org, anthropic.ml.
3. Deeply contemplate why you are racing humanity ahead, against our wishes. According to 377,458 people across 104 countries, 60% of the world wants development slowed, paused or stopped: niradata.com/ai-sentiment. And 30+ heads of state have said similar: presidentti.fi/en/a-call-for….
@hilbertspaess, @DKokotajlo, @dgrobinson, and the other amazing whistleblowers who have already come forth, please consider sharing your thoughts.
President Trump and President Xi could make a deal that ends the AI race before AI ends us.
It would be the most win-win deal of all time. A Nobel Prize-worthy achievement.
Sign and share to show your support: TrumpXiDeal.ai.
As has been long predicted, the world is about to discover how insecure its digital systems are against AI-augmented hackers. And AIs themselves.
Humanity will have to massively refocus on our collective existential safety in order to get through this.
To the cybersecurity heroes hard at work, we salute you.
To everyone else, please harden your digital systems and help with safety efforts: existentialsafety.org.
Last week some of South Korea's biggest banks were hit by a cyberattack. Thanks to a report from CrowdStrike tonight, we now know the entire hack may have been done by a single person. He used a combined stack of an open-source AI penetration tool named ARTEX, DeepSeek v4.1-Flash, GLM-5.3, Grok 4.6, and Claude Code.
"To manage risks from advanced AI systems we need clear safety standards, continuous testing, and reporting of serious failures. America can lead the world in AI by building systems that are not only more capable, but safer and more secure."
-@SenatorCantwell
It's extremely enheartening to see such a commonsense AI governance framework proposed for the US.
It's far past time to enact serious governance of the most dangerous technology ever invented.
.@SenatorCantwell outlined a comprehensive #AI governance framework built upon six principles that must drive the legislative action needed to address catastrophic risks and other AI harms:🧵
American voters should also support the Ban Artificial Superintelligence Act: xitter.cf/aisafetyaction/status/….
And encourage President Trump and President Xi to make a win-win deal to protect and benefit all people: xitter.cf/aisafetyaction/status/….
President Trump and President Xi could make a deal that ends the AI race before AI ends us.
It would be the most win-win deal of all time. A Nobel Prize-worthy achievement.
Sign and share to show your support: TrumpXiDeal.ai.
Per the "U.S. Led Global Cooperation in Adopting and Upholding Common AI Safety and Security Standards" principle, we strongly agree.
We propose international governance of AI, coupled with benefit-sharing for all people. Trump and Xi could shake on a win-win deal to make this a reality at any time: TrumpXiDeal.ai.
Per the "U.S. Led Global Cooperation in Adopting and Upholding Common AI Safety and Security Standards" principle, we strongly agree.
We propose international governance of AI, coupled with benefit-sharing for all people. Trump and Xi could shake on a win-win deal to make this a reality at any time: TrumpXiDeal.ai.
Per the "U.S. Led Global Cooperation in Adopting and Upholding Common AI Safety and Security Standards" principle, we strongly agree.
We propose international governance of AI, coupled with benefit-sharing for all people. Trump and Xi could shake on a win-win deal to make this a reality at any time: TrumpXiDeal.ai.
The policies that would help reduce catastrophic risks are clear.
And the political will to enact them is growing, thanks to ordinary people talking to their representatives. Keep at it!
AI experts largely agree about which policies would most reduce catastrophic AI risk.
The Longitudinal Expert AI Panel (LEAP) brings together forecasts from a representative group of over 250 experts that policymakers look to when considering the future of AI: computer scientists, industry figures, top economists, policy experts, and AI risk experts.
We asked panelists to forecast (1) how a menu of policies would affect AI risk; (2) which policies they each support; and (3) which policies they think will be implemented by 2030.
Key findings:
• Despite disagreeing on the level of AI risk, respondents broadly agree on policy
• Respondents believe government policy can substantially reduce catastrophic AI risk
• The most supported policy of those we asked about was the creation of an international body, including the US and China as members, with pre-release authorization power over frontier models
• Some US-only policies are forecast to reduce catastrophic AI risk, but not as much as bilateral policies involving the US and China
• Experts strongly oppose federal preemption and think it would increase AI risk
• The policies experts expect to reduce risk the most are those they think are least likely to actually be implemented
🧵 More, with key insights below:
