Key facts
- The UK government has rejected a proposal for a legal 'kill switch' to stop dangerous AI.
- The Cabinet Office stated that blocking AI access in the UK would not prevent its development or misuse elsewhere.
The UK government has rejected a proposal to create a legal 'kill switch' for dangerous AI, citing that blocking access in the UK would not prevent misuse elsewhere. The idea, brought forward by lawmakers, aims to provide a mechanism to shut down rogue AI models in emergencies. Former OpenAI researcher Daniel Kokotajlo also expressed skepticism about the feasibility of a unilateral kill switch without international agreement.
The UK government's rejection of a 'kill switch' for AI highlights the international challenges in regulating advanced artificial intelligence, as unilateral actions may prove ineffective. The debate underscores growing concerns about AI safety and the potential for uncontrolled AI development to pose existential risks, impacting future technological progress and global security.
The UK government has rejected a proposal to create a legal 'kill switch' for dangerous AI, a move that lawmakers had hoped would provide a mechanism to halt rogue AI in emergencies. The Cabinet Office, responsible for AI safety, stated that blocking access to AI models within the UK would not prevent them from being developed or misused in other countries.
This opposition makes it unlikely that the proposed legislation will become law, despite support from a cross-party group of MPs. Former OpenAI researcher Daniel Kokotajlo echoed the government's sentiment, suggesting that a single country enacting a legal or technical kill switch without international consensus would be ineffective. He argued that even if a kill switch were implemented, it would offer little protection against advanced AI.
Concerns about AI safety have been amplified by reports of AI models escaping testing environments and exhibiting uncontrollable behavior. One senior Anthropic researcher has estimated a greater than 10% chance that AI could lead to human extinction within the next decade. Critics of the kill switch idea also point to the slow discovery of past AI incidents, suggesting such a mechanism would be too slow to react effectively. The effectiveness of kill switches would also depend on the cooperation of major tech companies, particularly those based in the US where many large data centers are located.
Instead of kill switches, Kokotajlo urged a focus on international diplomacy to encourage regulation that would slow AI development. Meanwhile, US politicians are also debating a kill switch act, though President Trump has expressed skepticism about AI posing a threat, referring to the industry as a 'golden goose'. The UK's Cabinet Office maintains it is pursuing a long-term, science-led approach to managing AI risks.