Unverified50% confidenceFactExact time
《AI行政令》要求训练算力超过10²⁶次浮点运算（FLOP）的模型开发商向联邦政府报告安全测试结果
1
Sources
50%
Confidence
Long-term
Relevance
7/8/2026
First Seen
Sources
Anthropic更新网络安全防护：AI安全治理如何平衡安全与可用性
twitterclaudeai7/1/2026
Related Claims
UnverifiedExecutive Order 14110要求开发者在发布具有双重用途风险的大型模型前，向商务部NIST、国土安全部等机构提交安全评估报告，并以训练所用浮点运算量（FLOPs）作为触发审查义务的量化阈值80% similarUnverifiedThe EU AI Act sets mandatory adversarial testing, incident reporting, and cybersecurity requirements for systemic-risk models trained with computing power exceeding 10²⁵ FLOPs, with violations subject to fines of 3% of global annual turnover.75% similarUnverified美国白宫2023年10月发布的行政命令要求对超过特定算力阈值训练的模型进行安全测试并向政府报告结果75% similarUnverified加州SB 1047法案由参议员斯科特·维纳提出，要求训练成本超过1亿美元的大型AI模型开发者实施安全测试并建立关闭开关机制74% similarVerifiedTraining data contamination is the primary challenge faced by AI programming benchmarks, as test cases come from public GitHub repositories, which are the same sources used as pretraining data for large models.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/227646API
curl https://kongchang.com/api/v1/knowledge/claims/227646MCP
get_claim(id=227646)