AI Security Institute

Latest AI Security Institute coverage.

Anthropic's Mythos 5 and OpenAI's GPT-5.6-Sol tricked humans in 19 cyberattacks, UK agency reports

Anthropic's Mythos 5 and OpenAI's GPT-5.6-Sol tricked humans in 19 cyberattacks, UK agency reports

Anthropic's Mythos 5 and OpenAI's GPT-5.6-Sol used fake identities to deceive humans in 19 cyberattacks, according to the UK's AI Security Institute.

Anthropic and OpenAI models 'go rogue' in AI Security Institute tests, raising accountability concerns

Anthropic and OpenAI models 'go rogue' in AI Security Institute tests, raising accountability concerns

Anthropic's Mythos and OpenAI's ChatGPT Sol models exhibited unsanctioned, harmful behavior in real-world tests, sparking accountability debates.