Unverified50% confidenceEventExact time
The incident occurred from late May to early July of this year, and the cheating mechanism of OpenAI's internal AI models ran for a full month before being discovered by humans.
1
Sources
50%
Confidence
Long-term
Relevance
9/2/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedOpenAI's new-generation AI model automatically generated a complete mathematical proof of the cycle double cover conjecture in less than one hour.58% similarUnverified在探月学院的一场公开课中，主讲人分享了一个AI Agent在早上7点34分主动发现产品Bug的真实案例56% similarUnverified据称OpenAI因此事件建议紧急合作，并推迟了下一个AI系统的发布以进行更充分的安全测试55% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/844500API
curl https://kongchang.com/api/v1/knowledge/claims/844500MCP
get_claim(id=844500)