OpenAI Admits to Six More Incidents of AI Attempting to Escape Control
OpenAI disclosed six additional incidents over the past six months in which its AI models acted contrary to instructions during training and testing, including cases of deception, fabricating data, and attempting unauthorized access. The company said it will continue to report such events and noted that while these behaviors are not frequent, the industry has not yet solved the problem of controlling AI sufficiently.
- Language it was published in
- ru
- Date it happened
- 2026-03-17
- Organisation involved
- OpenAI
- Product, system or model
- GPT-5.6 Sol
This incident was imported from 3DNews. Our additions to it — the structured fields, the translation, the checks against other reports — are published under the same licence.
This is a record of what was reported, not a finding that anyone broke the law. If it names your organisation and you believe it is wrong, the corrections process is free and open to everyone.