Microsoft President Brad Smith stated that the United Kingdom is currently the most suitable nation to spearhead global artificial intelligence safety regulations. He argues that the U.K. possesses a unique position compared to the United States and the European Union, which allows it to implement oversight more effectively.
The British AI Security Institute, established in 2023, currently receives access to new models from major U.S. frontier labs for safety testing prior to their deployment. Despite recent pressure from U.S. President Donald Trump, who reportedly requested that companies like OpenAI and Anthropic withhold models from British testers, the institute confirmed it retains access to the latest technology.
Smith advocates for mandatory legal frameworks rather than voluntary self-regulation, noting that current standards for consumer goods are stricter than those for AI firms. He proposed specific measures, including legal requirements to report system breaches by rogue AI agents and mandatory approval from external evaluators before releasing new AI products.
The Microsoft executive also expressed support for implementing kill-switches to pause or slow down AI systems if human control is lost. These comments follow a meeting at the White House last month, where leaders from companies including OpenAI, Anthropic, Nvidia, AMD, SpaceX, Meta, and Google signed a document agreeing to work with independent auditors on model safety.
