Government officials, including Rishi Sunak’s advisers, are engaging in discussions with global leaders to formulate a formal statement addressing the risks associated with artificial intelligence (AI) before the upcoming AI Safety Summit scheduled for November 1st and 2nd. They aim to emphasize the UK’s leadership in AI safety, with a proposed domestic AI taskforce taking on a global role.

A draft agenda mentions the potential establishment of an “AI Safety Institute” focused on enabling national security agencies to evaluate advanced AI models effectively. However, the government had previously downplayed this idea. The Summit is not intended to create a new international institution but will concentrate on international collaboration.

The Summit’s objectives include updating safety guidelines previously released by the White House, exploring global cooperation on AI-related risks, and concluding with discussions among like-minded countries on how to scrutinize AI from a national security perspective.

This event will have a limited audience, with around 100 attendees, including cabinet ministers, CEOs of AI companies, academics, and representatives from international civil society. Prominent companies like OpenAI, Google, and Microsoft are expected to participate and disclose their adherence to AI safety commitments made in collaboration with the Biden administration in July. These commitments involve pre-release testing of AI models and ongoing scrutiny of their operation to ensure safety.