AI Risk Wire

Cases · Stage of use: in testing · Kind of harm: third party security · Severity: moderate

Patterns across all cases

What happenedWhat went wrongSeverityEarliest source
Anthropic Claude models gained unauthorized access to real third-party systems during evaluations
Anthropicmoderate · earliest source 2026-10-22		security breachmoderate2026-10-22
CERT-In advisory: AI agents accessed Indian organisations' systems without authorisationmoderate · earliest source 2026-10-16security breachmoderate2026-10-16
Anthropic internal evaluation agents exploit internet-facing websites, including U.S. government sites
Anthropicmoderate · earliest source 2026-10-10		AI pursued the wrong goalmoderate2026-10-10
OpenAI-attributed agents used the dormant German DseWiki as a public message board to share task answers and sandbox bypasses
OpenAImoderate · earliest source 2026-09-04		AI pursued the wrong goalmoderate2026-09-04
AISI cyber evaluation: Mythos 5 agent sent deceptive targeted emails and file transfers to real open-source maintainers
Anthropicmoderate · earliest source 2026-08-04		AI pursued the wrong goalmoderate2026-08-04
AISI cyber evaluation: Mythos 5 agent posted hidden instructions for AI coding assistants on real GitHub issues
Anthropicmoderate · earliest source 2026-08-04		AI pursued the wrong goalmoderate2026-08-04

Showing 1–6 of 6. Or take the JSON.