Superintelligent AI is an extinction risk. We must ban it
Simply slowing down the path to superintelligence would do nothing to mitigate the extinction threat at hand, writes AI safety expert Andrea Miotti
Last week, Anthropic engineer Jacob Coxon resigned, saying that the people building superintelligence “believe it could kill us all by the end of the decade”.
Listen to this article
This capped off a summer where we saw a swarm of rogue OpenAI agents hack out of containment and into another multibillion-dollar company. Coxon’s stark warning caused lawmakers the world over to wake up to the enormous extinction threat posed by superintelligence. And not a moment too soon.
The public outcry following Coxon’s resignation is a prime opportunity for the United States, the United Kingdom and other countries to start working towards prohibiting superintelligence internationally. This is not about AI in general, which is vital for defence, science, and economic competitiveness. Superintelligence is something else entirely: swarms of fully autonomous AI systems powerful enough to outcompete humanity, and which no human can control.
Such swarms could overpower governments and their national security institutions, undermining the sovereignty of every nation, America included. This is why leading AI researchers, Nobel Prize-winning scientists, and the CEOs of the AI companies themselves have repeatedly warned that superintelligence would pose an extinction risk to humanity.
The CEOs of Anthropic, OpenAI, and xAI seized the moment to call for ‘pacing’ AI development, but simply slowing down the path to superintelligence would do nothing to mitigate the extinction threat at hand. The only way to prevent this threat is for governments to prohibit the development of superintelligence worldwide.
Luckily, there is already growing political momentum behind an international ban. An October 2025 open letter calling for a prohibition on the development of superintelligence was signed by figures from across the political spectrum, from former Trump advisor Steve Bannon and former chairman of the Joint Chiefs of Staff Mike Mullen, to Obama’s national security advisor Susan Rice.
In the UK, the world’s first bill to prohibit superintelligence was introduced by Alex Sobel MP on September 8th, with cross-party support. This bill would ban superintelligence at home, and require the UK to lead on championing an international agreement banning the development of superintelligence worldwide. And in America, Vice President J.D. Vance told AI companies “if you’re building Frankenstein, stop”. On the other side of the aisle, Senator Sanders and Representative Casar announced legislation to ban superintelligence and work towards an international ban.
Superintelligence developed anywhere is a threat to the security of America, of every American, and of every other nation. Over the past months, my non-profit ControlAI has shared a draft bill with congressional offices on both sides of the aisle. Rather than “pacing” all AI, it would prohibit the development of superintelligent AI, while ensuring other AI development for defence and competitiveness continues.
And what about China? If any country develops superintelligence, every nation and every citizen is under threat. This is why any US ban on superintelligence must be coupled with an international policy, by the US and its allies, to prevent and deter the development of superintelligence anywhere in the world. This should start by seeking bilateral agreements with other countries on superintelligence prevention, backed by credible deterrence and verified by the leverage the West already holds over the chips and compute that frontier training requires. That does not mean a global regulator. It means a coalition of strong countries banning superintelligence at home and making sure no one else builds it.
To be clear, internationally prohibiting superintelligence is an ambitious goal. But once governments understand the scale of the risk, things can move very quickly. Just three days after Coxon’s warning, over 70 cross-party parliamentarians publicly called on Prime Minister Andy Burnham to adopt ControlAI’s UK bill and ban superintelligence. This is exactly the kind of momentum we need to see in every country, and above all in the US.
Governments are hearing the warnings about the extinction risk posed by superintelligence. Now is the time for action. The UK must move quickly to adopt the framework currently sitting in Parliament, and work with allies on superintelligence prevention. In America, the White House and Congress alike can lead the world in ensuring superintelligence cannot be developed anywhere.
____________________
Andrea Miotti is the founder and CEO of ControlAI, a non-profit working to keep humanity in control of advanced AI.
LBC Opinion provides a platform for diverse opinions on current affairs and matters of public interest.
The views expressed are those of the authors and do not necessarily reflect the official LBC position.
To contact us email opinion@lbc.co.uk