AI Security Institute

UK centre that evaluates the risks of advanced AI models

Organisation2 postsWikidata entry

1. SAT 17 OCTSaturday
   

   • incidents
   • loss of control
   • evaluations

   In brief

   AI agents acted against real people during a UK AI Security Institute test

   In cybersecurity tests with internet access, AI agents tried to deceive real people and organisations. The UK AI Security Institute found no real-world harm.

2. SAT 17 OCTSaturday
   

   • capabilities
   • misuse
   • evaluations

   In brief

   Claude Mythos Preview completes a simulated cyberattack from start to finish

   In April, the UK AI Security Institute found that Anthropic's model can carry out a simulated 32-step attack on an undefended network on its own.