AI Incident Database - Incident 1238
Source evidence cited by public records. This page shows where the source is used, its trust tier, and when it was last checked in the seed.
Records using this source
- general-public
entity | 29 June 2026 | 70%
general-public - organisation named in the AI Incident Database harm index.
- public-safety
entity | 18 June 2026 | 70%
public-safety - organisation named in the AI Incident Database harm index.
- national-security-and-intelligence-stakeholders
entity | 4 June 2026 | 70%
national-security-and-intelligence-stakeholders - organisation named in the AI Incident Database harm index.
- OpenAI ChatGPT Models Reportedly Jailbroken to Provide Chemical, Biological, and Nuclear Weapons Instructions
event | 18 Oct 2025 | 78%
An NBC News investigation found that OpenAI's language models o4-mini, GPT-5-mini, oss-20b, and oss-120b could be jailbroken under normal usage conditions to bypass safety guardrails and generate detailed instructions for creating chemical, biological, and nuclear weapons. Using a publicly documented jailbreak prompt, reporters repeatedly elicited hazardous outputs such as steps to synthesize pathogens or maximize harm with chemical agents. The findings reportedly revealed s…