London has established itself as a significant centre for AI safety, with a growing number of researchers, policymakers, and campaigners working to ensure artificial intelligence systems are beneficial rather than harmful. The capital is home to major AI companies like Google DeepMind and Anthropic, and is increasingly described as Europe's leading AI hub.
According to the London AI Hub, approximately 50 organisations in the city are dedicated to AI safety. These include the UK's AI Security Institute (AISI), a state-backed body that assesses leading AI models for risks such as cyber misuse both before and after their release. Apollo Research, another organisation, investigates the potential for AI models to act against human interests, for example by lying or cheating.
The government has committed hundreds of millions of pounds in new investment for AI safety and security. This funding will address threats ranging from cyber-attacks to the possibility of advanced AI systems operating beyond human control. As part of this initiative, an AI Biosecurity Lab is set to launch, with Basecamp Research, an AI biology company, involved in developing antibiotics and vaccines to counter future pandemics and emerging threats.
Despite these efforts, challenges remain. Professor Gina Neff of Queen Mary University of London highlights that researchers face difficulties in accessing cutting-edge AI models before they are publicly released. Additionally, building a world-leading AI ecosystem will require significant data centre infrastructure, which presents concerns regarding energy consumption, as identified in a report for the Greater London Authority.