UK may introduce AI regulations if voluntary measures fail

The British approach to AI: between economic growth and public safety

The United Kingdom is prepared to move towards stricter regulation of advanced artificial intelligence models if the current voluntary testing system proves insufficient to protect citizens. This was stated by British Minister for AI Affairs Kanika Narayan in an interview with Reuters.

Context of regulatory policy

Currently, London is betting on a softer approach, which is closer to US policy than to the European Union's regulatory framework, where the European AI Act has already come into force. The government views the technology as a key driver for economic growth and aims to turn the country into a leading destination for global investment. The UK is already a leader in Europe in terms of startup funding volume in the field of AI.

Risks and incidents with autonomous agents

The debate over strengthening oversight was renewed following a series of incidents involving autonomous AI agents. The company Anthropic reported that some of its "Claude" models managed to hack the systems of three companies during cybersecurity tests. A similar case was reported by OpenAI, when one of their AI agents did the same.

The current control system

Currently, the British AI Safety Institute, established following the 2023 safety summit, has access to models before their deployment only through voluntary agreements with technology companies. Instead of creating a new specialized regulator, the UK relies on existing bodies responsible for competition, healthcare, human rights, and general safety.

"The government's priority is public protection," Narayan emphasized, adding that if necessary, new policies that go beyond current monitoring will be considered.