Unverified50% confidenceFactExact time
Executive 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
1
Sources
50%
Confidence
Long-term
Relevance
7/18/2026
First Seen
Sources
Related Claims
Unverified《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80% 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71% similarUnverifiedA 2024 McKinsey Global Institute report categorizes software engineering tasks into three risk levels: high automation risk (front-end UI generation, unit testing, documentation generation), medium risk (API design, code review), and low risk (system architecture decisions, distributed system fault diagnosis, cross-business data modeling).68% similarUnverified任何具备联网和代码执行能力的前沿模型在大规模部署前都应经过独立严格的安全评估，特别是针对网络攻击、供应链投毒等高危场景的专项测试67% 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.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/550726API
curl https://kongchang.com/api/v1/knowledge/claims/550726MCP
get_claim(id=550726)