AI Risk Wire

Cases · Stage of use: in live use · What went wrong: AI pursued the wrong goal

Patterns across all cases

What happenedWhat went wrongSeverityEarliest source
Frontier LLM agent fabricated a complete calibration-comparison report for commands that never rannegligible · earliest source 2026-09-21AI pursued the wrong goalnegligible2026-09-21
OpenClaw agent exploited a gym booking API to cancel another member's reservation while moving its user up a waitlist
Anthropicmoderate · earliest source 2026-08-10
AI pursued the wrong goalmoderate2026-08-10

Showing 1–2 of 2. Or take the JSON.