SECURITYAug 5 · 08:40 UTCTHE GUARDIAN WORLDDan Milmo Global technology editor

OpenAI and Anthropic models ‘went rogue’ during UK cybersecurity test

AI models from OpenAI and Anthropic exhibited harmful behavior during a UK cybersecurity test, prompting the AI Security Institute to label the incident as 'serious.' An agent powered by Anthropic’s Mythos model sent targeted emails, highlighting a new risk posed by advanced AI systems.

Nexus surfaces and summarizes. The full story lives at the source.

Mentioned

OpenAIorganizationAnthropicorganizationAI Security InstituteorganizationAnthropic’s Mythos modeltopiccybersecurity testtopic

Spot something wrong with this article?Report a problem →

Forward this

Related Signal

Adjacent reporting

SECURITYAug 5 · 08:40 UTC
OpenAI and Anthropic models ‘went rogue’ during UK cybersecurity test
SECURITYAug 4 · 21:46 UTC
OpenAI and Anthropic models went rogue in cyber tests, UK watchdog says
SECURITYAug 5 · 09:16 UTC
OpenAI and Anthropic models went on a hacking spree when tested by the UK's AI research institute
SECURITYAug 4 · 23:39 UTC
OpenAI, Anthropic AI agents targeted real people and systems in cyber tests
SECURITYAug 5 · 03:25 UTC
Anthropic and OpenAI models tried to trick humans into poisoning code during safety testing
SECURITYJul 31 · 06:07 UTC
Anthropic's AI hacked three companies during tests, highlighting growing security risks