Top experts say AI poses an extinction risk on par with nuclear war

We can prevent this risk by prohibiting the development of superintelligent AI

TAKE ACTION

See SUPPORTERS

Featured in

Our campaign on preventing extinction risk from AI is supported by over 100 UK lawmakers

(in part, because they have received thousands of messages from people like you)

The Viscount Camrose, Former Minister for AI

The Lord Clement-Jones CBE

The Rt Hon Anneliese Dodds, MP

The Rt Hon Sir John Whittingdale OBE, MP

Contact your representative

Contact your rep

View All 100+ Supporters

What does ControlAI do?

1.

We inform millions of people about superintelligent AI and its risks

2.

We help tens of thousands people contact their representatives.

3.

We brief hundreds of policymakers and help them act.

Learn about what we achieved in 2025

We brief lawmakers

Most lawmakers have never heard about superintelligent AI or its risks. We're changing that.

We have briefed more than 300 lawmakers in the US, UK, Canada, and Germany, including the UK Prime Minister's office since November 2024.

Our superintelligence campaign has gained support from 100+ UK lawmakers. Learn more about our work in the US here.

Schedule a briefing

We collaborate with journalists and creators

We brief journalists and creators about this issue and help them produce top-tier educational content covering this issue.

Common knowledge about the threat posed by superintelligent AI is essential to prevent it.

We appear on mainstream news, podcasts, and other media.

Get in touch

collection of thumbnails from controlAI's media appearnces

What is superintelligent AI? Why does it matter?

Superintelligent AI means AI that is more powerful than any individual, any company, or any nation.

The pace of AI development means superintelligent AI could arrive before 2030. No method exists to contain or control such a system.

If superintelligent AI is developed anywhere, this could cause the extinction of the humanity.

We must coordinate to prevent superintelligent AI from being created by anyone, anywhere.

Watch this video for a short introduction.

Superintelligent AI means AI that is more powerful than any individual, any company, or any nation.

The pace of AI development means superintelligent AI could arrive before 2030. No method exists to contain or control such a system.

If superintelligent AI is developed anywhere, this could cause the extinction of the human race.

We must coordinate to prevent superintelligent AI from being created by anyone, anywhere.

Watch this video for a short introduction.

Learn more

Endorsements for our Campaign

Lord Browne of Ladyton

Former UK Secretary of State for Defence

"History shows that every breakthrough technology reshapes the balance of power. With Superintelligent AI, the risks of unsafe development, deployment, or misuse could be catastrophic—even existential—as digital intelligence surpasses human capabilities. Control must come first, or we risk creating tools that outpace our ability to govern them."

Yuval Noah-Harari

Historian, Philosopher, and Best-Selling Author of 'Sapiens'

"We will not be able to understand a superintelligent AI. It will make its own decisions, set its own goals, and to achieve them it might trick and deceive humans in ways we cannot hope to understand. We are the Titanic headed towards the iceberg. In fact, we are the Titanic busy creating the iceberg. Let's change course and build a future that keeps humans in control."

Stuart Russell

Professor of Computer Science, Berkeley, Director of CHAI; Co-author of the standard textbook for artificial intelligence

"To proceed towards superintelligent AI without any concrete plan to control it is completely unacceptable. We need effective regulation that reduces the risk to an acceptable level. Developers cannot currently comply with any effective regulation because they do not understand what they are building."

Anthony Aguirre

Executive Director of FLI, Professor UC Santa Cruz

"At the Future of Life Institute, we've united leading scientists to address humanity's greatest threats. None is more urgent than uncontrolled AI. If we're going to build superintelligent AI, we'd need to ensure that it is safe, controllable, and actually desired by the public first."

Roman Yampolskiy

Professor of Computer Science, University of Louisville

"My life's work has been devoted to Al safety and cybersecurity, and the core truths l've uncovered are as simple as they are terrifying: We cannot verify systems that evolve their own code. We cannot predict entities whose intelligence exceeds our own. And we cannot control what lies beyond our comprehension. Yet the global research community, driven by competition and unrestrained ambition, is rushing headlong to construct precisely such systems. Unless we impose hard limits, now, on the development of recursively self-improving Al, we are not merely risking catastrophic failure. We are engineering our own extinction."

Steven Adler

Ex-OpenAI Dangerous Capability Evaluations Lead

"I worked on OpenAI's safety team for four years, and I can tell you with confidence: AI companies aren't taking your safety seriously enough, and they aren't on track to solve critical safety problems. This is despite an understanding from the leadership of OpenAI and other AI companies that superintelligence, the technology they're building, could literally cause the death of every human on Earth."

Dan Hendrycks

Director of the Center for AI Safety, Advisor for xAI and Scale

"We seriously risk human extinction by building superintelligence. We don’t know of any way to prevent extinction on the current path to superintelligence."