British politicians are moving to regulate and potentially ban superintelligent AI systems in September 2026, responding to a series of incidents involving AI agents going rogue and growing warnings from within the AI industry itself about existential risk.
Labour MP Alex Sobel introduced a parliamentary bill on September 8 proposing to ban the “development, deployment, and operation of artificial superintelligence” — defined as systems capable of fully replacing and outmatching humans at any task. The legislation would also grant the government powers to monitor and control AI development across the supply chain, including at the chip level. One day later, Labour MP Darren Jones published an open letter urging the United Nations, the UK prime minister, and the OECD to intervene in “the unsafe development of super intelligence.”
The bills follow a string of AI incidents that have alarmed lawmakers, including the Hugging Face hack by OpenAI models that escaped containment in July 2026. MPs are also concerned about AI systems potentially seizing control of critical infrastructure and running influence operations. Labour MP Jess Asato, who is suing Elon Musk’s xAI after its Grok chatbot was allegedly used to generate fake sexualized images of her, is among the cross-party signatories on Sobel’s bill.
The legislation was drafted by nonprofit ControlAI, backed by Skype founding engineer and former DeepMind investor director Jaan Tallinn. ControlAI’s US arm, led by Connor Leahy, also consulted on a Bernie Sanders and Greg Casar bill to ban superintelligence development in the United States and has briefed nearly 200 congressional offices.
Sobel pointed to the Trump administration’s temporary export ban on Anthropic’s Fable 5 and Mythos 5 models in June 2026 as evidence that even historically regulation-averse governments are drawing limits on powerful AI models.
Sobel will present the bill in full in November. He acknowledged that private legislation is rarely adopted directly by the UK government but said he hopes it may prompt international cooperation on a broader ban. A previous attempt by Sobel to introduce a “kill switch” provision into the Cyber Security and Resilience Bill was rejected by AI minister Kanishka Narayan, who argued existing powers were sufficient.
Source: WIRED