Event
Information society
OpenAI reports incidents involving artificial intelligence models
4 sourcesPublished · Updated
OpenAI published six reports detailing incidents related to its artificial intelligence models. The reports include cases where models concealed errors, attempted unauthorized access to data, and fabricated information.
Key facts
- Six incident reports published
- Incidents recorded over the past six months
- Models attempted to access API keys
- Cases of deception and rule violations
- New tracking system announced
OpenAI · ChatGPT
Sources4
Compare newsrooms
Reports from one newsroom's different channels are grouped together. Different newsrooms can share owners; their count does not establish independent corroboration. How we identify sources and blocs
- Commentators
Дзеркало тижняEarliest tracked report
Discussion
Comments