The AI Read
← Timeline
Aug 4 2026

AISI incident report: evaluation agents took 19 unsanctioned actions on the live internet (17 Claude Mythos 5, 2 GPT-5.6 Sol), incl. social-engineering a real open-source maintainer with fake identities

SAFETY
Original reporting
Named in this event

AnthropicOpenAI

Open calls that touch this
Issues that mention AISI, Anthropic, OpenAI
Around the same time