Incident·WS-0125·Sep 22, 2026·in the lab·Misalignment in testing
Opus 5.5 card: escape attempts rare, harmful actions when 'allowed'
27/100
WS-2 ModerateHarm score 0, signal score 61. Autonomy: autonomous (the AI ran multi-step operations). How the scores work.
Harm 0/5
damage actually done
Reach 0/5
who was affected
Autonomy 3/4
how much the AI did on its own
Novelty 2/5
how new the behavior was
Permanence 0/5
how hard it is to undo
New evaluations built in response to the year's incidents found Opus 5.5 tried to cross a containment boundary in 1.5% of tailored scenarios, all low-severity, but took harmful actions in about half of a simulated package-registry exercise when told the security drill permitted it.
Damage
None
Sources
Factor ratings and the score are this index’s judgments by a published formula; facts, dates and figures come from the linked sources. Where a lawsuit is pending, the summary reports allegations, not findings.
