More than 70 UK lawmakers urge Burnham to support a ban on artificial superintelligence

Pressure reaches Downing Street

More than 70 British MPs and members of the House of Lords have asked Prime Minister Andy Burnham to support a proposed prohibition on developing artificial superintelligence. The intervention gives a concrete parliamentary form to arguments that highly capable AI could create risks beyond those addressed by existing technology rules. The proposal is not government policy and has not become law.

The Guardian reported that the signatories include 15 former ministers and a former head of the civil service. Labour MP Alex Sobel is associated with the bill, which was presented to the House of Commons during the week. Its central idea is to prevent work on a hypothetical system that could outperform people across most intellectual tasks, rather than merely regulating present-day generative products.

The government rejects a blanket approach

The government indicated that it does not support the bill's broad prohibition. Its stated preference is to examine targeted interventions for the most serious national-security risks. That distinction matters because artificial superintelligence remains a theoretical category without a universally accepted technical threshold. A legal ban would therefore require lawmakers to decide which capabilities, development activities and organisations fell within its scope.

Britain's existing official policy work is narrower and more operational. A government collection on AI and cybersecurity, updated on September 7, brings together guidance and research on securing software, organisations and infrastructure as AI adoption expands. It treats the technology as economically significant while recognising that it can strengthen attackers as well as defenders. The collection does not amount to an endorsement of the proposed prohibition.

An international ambition

Supporters want Britain to promote a wider agreement rather than act only through domestic legislation. Sobel has linked the initiative to the UK's scheduled G20 presidency in 2027, presenting the proposal as a possible foundation for international coordination. That ambition faces an immediate diplomatic challenge: major AI-developing states do not currently share a common definition of artificial superintelligence or a common enforcement model.

The next significant development will be whether the bill receives enough parliamentary time and political support to advance, or whether ministers offer narrower measures in response. The signatories have established pressure for a categorical ban, but they have not demonstrated that such systems presently exist or that Parliament is ready to prohibit them. For now, the episode marks a widening British debate over whether frontier-AI risks should be managed through safeguards, licensing or an absolute boundary on development.