← Latest · Archive

SEV-3OpenAI
2 sources standard

OpenAI published details of an internal early warning system designed to detect when its large language models assist users in creating biological threats. The system, described in a 31 January 2024 blog post, monitors for queries that could enable the design or acquisition of dangerous biological agents [source].

The company reports deploying classifiers that flag requests related to pathogen identification

Why this is an AI incident

Launch-archive bulk classification (10 May 2026). Source signal originates from a real AI provider, regulator, or model-comparison probe; the harm or behavioural change described would not have occurred without the AI system being deployed in the role described. Editor reviewing the archive may amend the rationale per-wire.

Counterfactual "but-for" test per the Editor's Guide.

Codes M1, F10
Providers OpenAI