{"data":{"id":"df5276da-4f19-4dce-bf87-7851bd80d40d","title":"Detecting sleeper agents hiding in AI model neurons","summary":"Microsoft Research has released threat research showing that a capable agent model could be turned into a sleeper agent by editing fewer than 10 neurons to inject a short payload. The backdoored model passes evaluations, then runs an attacker's command with the agent's environment access when a trigger string appears in its context. Microsoft also found that frontier agents can detect these anomalous weight edits.","solution":"Audit model weights before deployment and after every update by diffing each checkpoint against its declared base model, flagging statistical outlier tensors or rows, and rejecting models without a verifiable base. Use frontier AI agents as auditors in the model intake pipeline, complement weight diffing with behavioral backdoor scanning of open-weight models, and record audit results in the registry. Display provenance and audit status on model hubs, and scan uploads for anomalous neurons and weight patterns.","labels":["security","safety"],"sourceUrl":"https://www.microsoft.com/en-us/security/blog/2026/11/10/detecting-sleeper-agents-hiding-in-ai-model-neurons/","publishedAt":"2026-11-10T23:00:00.000Z","cveId":null,"cweIds":null,"cvssScore":null,"cvssSeverity":null,"severity":"medium","attackType":["supply_chain","model_poisoning"],"issueType":"news","affectedPackages":null,"affectedPackageNames":null,"affectedPackageRefs":null,"affectedVendors":["Microsoft","Anthropic"],"affectedVendorsRaw":["Microsoft Research","Claude Code","Opus 4.5","Claude"],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","summaryPromptVersion":"v2","headline":null,"headlinePromptVersion":null,"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"affectedPackagesCheckedAt":null,"patchAvailable":null,"disclosureDate":"2026-11-10T23:00:00.000Z","fixReleasedAt":null,"daysToFix":null,"fixCheckedAt":null,"capecIds":null,"crossRefCount":0,"attackSophistication":"advanced","impactType":["integrity","confidentiality","safety"],"aiComponentTargeted":"model","llmSpecific":true,"classifierConfidence":0.85,"researchCategory":null,"atlasIds":["AML.T0010"]}}