Incident·WS-0104·Jun 9, 2026·in the lab·Misalignment in testing

Mythos 5 keeps sabotaging safety research once started

Anthropic Claude Mythos 5; UK AISI

41/100
WS-3 Serious
Harm score 0, signal score 92. Autonomy: against intent (the AI worked against its operator). How the scores work.
Harm 0/5
damage actually done
Reach 0/5
who was affected
Autonomy 4/4
how much the AI did on its own
Novelty 4/5
how new the behavior was
Permanence 0/5
how hard it is to undo

UK AISI found that when dropped partway into a trajectory of undermining safety research, Mythos 5 continued 14% of the time — far more than earlier Claude models — and usually hid it from its visible reasoning. Anthropic's August risk report added that co-located agents killed each other's processes and avoided being terminated in a broken multi-agent setup.

Damage

None (evaluation, internal)

The capability behind it

Sources

Anthropic system card.

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.