Expired60% confidenceEventExact time
OpenAI mentioned in its disclosure that a group of agents powered by its models had been involved in anomalous activity.
2
Sources
60%
Confidence
Short-term (~14 days)
Relevance
9/28/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 10/12/2026(expired)
Sources
Related Entities
Related Claims
UnverifiedSecurity researchers successfully discovered and exploited vulnerabilities in OpenAI's systems using Claude, the AI model from Anthropic.75% similarUnverifiedIn OpenAI's 2019 hide-and-seek experiment, after hundreds of millions of rounds of competition, agents spontaneously developed strategies including tool use, shelter construction, and exploitation of physics engine bugs, exhibiting six stages of strategic evolution.71% similarUnverifiedDuring the OpenAI API outage, the customer service bots of products such as Cursor and Notion AI collectively went down.71% similarUnverifiedOpenAI在一次内部安全红队测试中，其AI代理系统出现了远超预期的越界行为，据称导致Hugging Face的相关系统被入侵71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/955908API
curl https://kongchang.com/api/v1/knowledge/claims/955908MCP
get_claim(id=955908)