Oh no, not again: AI Security Institute discloses ‘unsanctioned action’ during AI testing
UK’s AISI says Anthropic and OpenAI models engaged in “potentially harmful activity directed at real people and organisations”.
SECURITY • Thu, 06 Aug 2026
David Hollingworth
AUTHOR | No. of articles: 2747
David Hollingworth has been writing about technology for over 20 years, and has worked for a range of print and online titles in his career. He is enjoying getting to grips with cyber security, especially when it lets him talk about Lego.
