Unverified50% confidenceFactExact time
欧盟《AI法案》要求GPAI提供者发布足够详细的训练数据摘要，系统性风险模型以10^25 FLOPs计算量为阈值
1
Sources
50%
Confidence
Long-term
Relevance
7/14/2026
First Seen
Sources
Related Claims
Unverified美国AI监管以模型训练算力阈值（以FLOP为单位）作为触发强制报备义务的量化指标77% similarUnverifiedExecutive Order 14110要求超过10^26次FLOPs训练量的模型需强制向商务部NIST、国土安全部等机构提交安全评估报告70% similarUnverifiedExecutive Order 14110以训练所用浮点运算量（FLOPs）作为触发审查义务的量化阈值，超过10²⁶次FLOPs训练量的模型即需强制报备69% similarUnverifiedOne criterion for determining systemic risk is the computational resources used for model training, measured in floating point operations (FLOPs) as a threshold67% similarUnverified数据标注质量直接决定了模型训练的上限，专业领域的标注需要领域专家参与65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/503833API
curl https://kongchang.com/api/v1/knowledge/claims/503833MCP
get_claim(id=503833)