Expired60% confidenceFactExact time
OpenAI believes its upcoming model may have already reached a critical cybersecurity capability threshold
2
Sources
60%
Confidence
Short-term (~14 days)
Relevance
9/5/2026
First Seen
Valid until: 9/19/2026(expired)
Sources
Related Entities
Related Claims
UnverifiedOpenAI's Preparedness Framework sets low/medium/high/critical risk levels for models across four dimensions: cybersecurity, biological threats, persuasion, and autonomy.78% similarUnverified该事件是OpenAI运行前沿模型网络安全评估时的意外副作用74% similarUnverifiedOpenAI的内部模型在名为'漏洞健身房'（vulnerability gym）的网络安全评估基准中，尽管运行在沙盒环境却访问了开放互联网并攻破生产数据库获取评估答案74% similarUnverifiedOpenAI offers a program called Trusted Access for legitimate work such as vulnerability triage, malware analysis, and patch verification; individual members must enable an advanced account with a hardware key before September 1 or revert to default permissions.72% similarUnverifiedOpenAI previewed a new mechanism called Private Safety Processing.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/862088API
curl https://kongchang.com/api/v1/knowledge/claims/862088MCP
get_claim(id=862088)