Edited by 
Short answer
The Statement on Superintelligence, organised by the Future of Life Institute and first published in October 2025, calls for "a prohibition on the development of superintelligence" until there is broad scientific consensus that it can be done safely and controllably, and strong public buy-in. Signatories include AI pioneers Geoffrey Hinton and Yoshua Bengio, Apple co-founder Steve Wozniak, Richard Branson, Prince Harry and Meghan, and figures from both the US left and right. A US poll released with it found 64% of Americans think superintelligence should not be built until it is proven safe. AI companies and many policymakers disagree, arguing a ban is unenforceable and would hand the lead to rivals.
What the statement says
The full statement is short. It notes that leading AI companies have stated a goal of building superintelligence in the coming decade, AI that would significantly outperform humans on essentially all cognitive tasks, and lists the concerns this raises: economic obsolescence and disempowerment, loss of freedom and dignity, national security risks and even human extinction. Its central line is:
"We call for a prohibition on the development of superintelligence, not lifted before there is 1) broad scientific consensus that it will be done safely and controllably, and 2) strong public buy-in."
It is not a call to stop AI in general. Today's chatbots, medical AI and coding tools are not covered. The target is the specific goal of building systems far smarter than humans. Read what is superintelligence and AGI vs ASI for the difference.
Who signed it?
The statement stands out for how broad its supporters are. Early signatories included:
- AI researchers: Turing Award winners Geoffrey Hinton and Yoshua Bengio, and Stuart Russell, author of the standard AI textbook.
- Business leaders: Steve Wozniak and Richard Branson.
- US national security figures: former National Security Advisor Susan Rice and retired Admiral Mike Mullen, former Chairman of the Joint Chiefs of Staff.
- Media and politics from both sides: Steve Bannon and Glenn Beck, alongside progressive figures.
- Public figures: Prince Harry and Meghan, Stephen Fry, Joseph Gordon-Levitt and will.i.am.
- Faith leaders and Nobel laureates, including economist Daron Acemoglu and physicist Frank Wilczek.
The list has kept growing since launch, because members of the public can add their names too.
What do Americans think?
Alongside the statement, the Future of Life Institute released a national US poll. Its headline findings:
| Position | Share of US adults |
|---|---|
| Support the current, largely unregulated approach | 5% |
| Support robust regulation of advanced AI | 73% |
| Say superintelligence should not be developed until it is proven safe and controllable | 64% |
Polls commissioned by advocacy groups should be read with care, but independent surveys by Pew Research Center and Gallup have also found that more Americans are concerned than excited about AI, and that large majorities want more oversight of it.
How does US policy treat superintelligence?
There is no US law that limits building superintelligence. Federal policy has focused on keeping the US ahead in AI while testing the most powerful models for national security risks:
- America's AI Action Plan (July 2025): the White House strategy prioritises AI innovation, data center and energy buildout, and exporting American AI, while removing rules seen as slowing development.
- Center for AI Standards and Innovation (CAISI): the renamed US AI Safety Institute at NIST evaluates frontier models, mainly for cyber, biological and other national security risks. In May 2026 it announced pre-deployment testing agreements with Google, Microsoft and xAI, and said it had completed more than 40 model evaluations.
- States: several states, led by California and New York, have passed laws requiring large AI developers to publish safety frameworks and report serious incidents.
Read AI safety in Australia to compare how another country is approaching the same question.
The case for a ban
Supporters make four main arguments:
1. No one knows how to control it. There is no proven method to make a system much smarter than its creators reliably follow human intentions. Building it first and solving safety later is, in their view, a gamble with no second chance. 2. Race dynamics make it worse. Companies and countries competing to be first have strong incentives to cut corners on safety. 3. The public has not agreed. A technology that could reshape every job and decision should not be built without democratic consent. 4. Prohibitions have worked before. Supporters point to international limits on biological weapons, human cloning and ozone-depleting chemicals.
The case against a ban
Critics, including most AI companies and many economists and policymakers, argue:
1. It cannot be enforced. There is no clear line between "advanced AI" and "superintelligence", and research happens in many countries. 2. It would hand the lead to rivals. If the US or its allies stop, others may not, and the first superintelligence might be built by the least careful developer. 3. It would block huge benefits. Superhuman AI could speed up medicine, science and clean energy. Read what superintelligence could do. 4. Safety can be built in along the way. Labs such as Safe Superintelligence (SSI), Anthropic, OpenAI and Google DeepMind argue that alignment research, testing and gradual deployment are a better path than a pause.
What a prohibition would mean for AI hardware
Any real limit would almost certainly target compute, because training frontier AI needs huge, visible clusters of chips. Proposals include licensing training runs above a certain amount of computing power, tracking where advanced GPUs go, and on-chip mechanisms to verify how chips are used. The US already restricts exports of top AI chips to some countries. Read what is an AI data center and what is a GPU cluster for the hardware such rules would cover.
Frequently asked questions
Is superintelligence banned anywhere?
No. As of November 2026, no country prohibits building superintelligence. The EU AI Act regulates high-risk uses and general-purpose AI models but does not ban research towards superintelligence.
Who wrote the superintelligence statement?
It was organised by the Future of Life Institute (FLI), a US-based nonprofit that works on risks from advanced technology. FLI also organised the 2023 open letter calling for a six-month pause on training AI systems more powerful than GPT-4.
Did AI company leaders sign it?
Most leaders of the major AI labs did not sign. Some have separately said that advanced AI poses serious risks while arguing that careful development, rather than a ban, is the right response.
Can I sign the statement?
Yes. The statement is open for public signatures on the Future of Life Institute's website.
Sources and further reading
- Future of Life Institute, Statement on Superintelligence
- Future of Life Institute, Press release and US polling
- Fortune, US administration and CAISI model testing (May 2026)
- NIST, Center for AI Standards and Innovation
Leave a Reply