Unverified50% confidenceEventExact time
2023年7月，OpenAI、Google、Anthropic等七大AI公司在白宫签署自愿承诺，将第三方安全测试列为核心条款
1
Sources
50%
Confidence
Long-term
Relevance
7/12/2026
First Seen
Sources
The Boundaries of AI Safety Guardrails: Deep Reflections Sparked by a Single Prompt
redditr/ChatGPT7/11/2026
Related Claims
UnverifiedThe White House requires that advanced models serving overseas users must be approved on a per-customer basis, and Trump signed an executive order stipulating that models must pass AI safety testing before release.63% similarUnverifiedExecutive Order 14110 signed in the United States in 2023 requires large AI models to submit safety test reports to the government before release62% similarUnverified2023年12月拜登政府的行政命令让微软、Anthropic、OpenAI、谷歌、亚马逊等七家公司承诺给模型加水印，但特朗普上任第一天该命令即被废除60% similarUnverified该自愿承诺协议的法理依据来自《国防生产法》第708条授权的行业协调机制，涵盖红队测试标准化、安全信息共享、AI生成内容水印标记等八项核心条款56% similarUnverifiedThe AI executive order EO 14110 issued by the Biden administration in 2023 required developers of dual-use foundation models to report safety test results to the government; the Trump administration revoked this executive order in 2025.56% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/490677API
curl https://kongchang.com/api/v1/knowledge/claims/490677MCP
get_claim(id=490677)