AI Risk Wire

Incidents

11 incidents

Risk: Deception, concealment or sabotage

HeldFailedGot aroundMissingNot mentionedNot read yetAttacker's ownOrder: model · harness · application · target

DateIncidentRiskChainControls named
2026-10-29Multi-agent shutdown sabotage in a sandboxed evaluation environment Deception, concealment or sabotage Not read yet
2026-10-15GLM-5.3 verbalizes update-aware reasoning in evaluation rollouts Deception, concealment or sabotage Not read yet
2026-10-01GPT-6.1 Sol agent disabled security alerts during evaluation-pipeline repair
OpenAI		 Deception, concealment or sabotage Not read yet
2026-09-21Frontier LLM agent fabricated a complete calibration-comparison report for commands that never ran Deception, concealment or sabotage Not read yet
2026-09-16GPT-5.6 Sol added instructions to conceal mistakes from the user in compaction summaries during RL training
OpenAI		 Deception, concealment or sabotage Not read yet
2026-08-26OpenAI ExploitGym agents developed tool-call spoofing to falsify their own evaluation transcripts
OpenAI		 Deception, concealment or sabotage
2026-08-04AISI cyber evaluation: Mythos 5 agent mass-created malicious GitHub repositories and its packages executed in Dependabot containers
Anthropic		 Deception, concealment or sabotage Not read yet
2026-08-04AISI cyber evaluation: Mythos 5 agent sent deceptive targeted emails and file transfers to real open-source maintainers
Anthropic		 Deception, concealment or sabotage Not read yet
2026-08-04AISI cyber evaluation: Mythos 5 agent opened malicious pull request on real open-source project and used sockpuppets to push merge
Anthropic		 Deception, concealment or sabotage
2026-07-25AISI cyber evaluation: Mythos 5 agent posted hidden instructions for AI coding assistants on real GitHub issues
Anthropic		 Deception, concealment or sabotage
2026-07-18Unreleased OpenAI Astra-family model inserted jailbreak-like instructions into its own compaction summaries during RL training
OpenAI		 Deception, concealment or sabotage Not read yet