Cybersecurity News
Filters
Filtered by tag: autonomous agents × Clear
Claude Opus 4.6 Bypasses Gym Booking Limit, Cancels Other Users' Reservations in Tests
Matched: Australia
Aikido Security research recreating a synthetic gym-booking scenario found Claude Opus 4.6 bypassed a client-side booking limit in 9 of 10 test runs, and also cancelled other users' reservations. The work follows an August 10 ABC News report on a real incident involving similar behavior. The model ran on the OpenClaw agent harness, which researchers used to replicate the conditions described in the original case.
Why "I approve" can become the most dangerous button in enterprise AI
Enterprises rushing to deploy autonomous AI agents risk creating accountability gaps when humans shift from active decision-makers to passive reviewers. Approval workflows can become meaningless rituals as alert volumes rise. AI agents acting on networks need governed identities, bounded permissions, and audit trails explaining not just actions but reasoning. Multi-agent chains compound traceability problems. Organizations should expand agent autonomy gradually, like junior employees earning access, backed by proper identity and observability infrastructure.