EU AI Act
The EU AI Act is a comprehensive regulatory framework established by the European Union to govern the development, deployment, and use of artificial intelligence systems within EU territory. The legislation classifies AI applications according to risk level, imposing transparency, safety, and compliance requirements on high-risk AI systems while explicitly prohibiting certain high-risk AI uses. It applies to all relevant entities operating in the EU market and is one of the world's first systematic legal frameworks for AI.
Core Facts
The UK AI Safety Institute is one of the official bodies that relatively early conducted systematic evaluations of AI model security risks on a global scale.
The EU Artificial Intelligence Act (AI Act) explicitly requires that high-risk AI systems must have human oversight and intervention capabilities.
欧盟的《人工智能法案》（EU AI Act）于2024年正式生效，将AI系统按风险等级分类管理
Timeline (last 90 days)
The EU AI Act has officially taken effect, requiring clear labeling of synthetic output and disclosure of bot identity to users, or otherwise facing hefty fines.
The EU AI Act and content moderation regulations across industries require that generated content be explainable and reviewable in specific scenarios.
Multiple countries and regions (including the EU AI Act) have classified automated recruitment decision-making as high-risk AI applications, requiring that a human final review step be retained.
Both the U.S. CISA and the EU AI Act list offensive cyber capabilities as a core evaluation dimension for high-risk AI systems.
OpenAI launched a text watermarking system called TextScreen, which embeds statistical signals in the model's word choices rather than inserting visible markers for subsequent detection, in response to the EU AI Act, and plans to open-source it.
OpenAI is planning to add watermarks to text generated by ChatGPT in order to meet EU regulatory requirements for AI content provenance.
The concept of frontier models has been adopted by regulatory bodies (such as the EU AI Act) as a category of models requiring special scrutiny.
Regulatory frameworks (such as U.S. SEC guidelines or the EU AI Act) generally require financial institutions to maintain explainability and auditability of automated decisions.
Under the EU AI Act and the U.S. NIST AI Risk Management Framework, a model's ability to refuse harmful instructions is regarded as a positive indicator of high safety standards, whereas the same capability may be characterized as an operational risk under defense procurement logic.
The EU's GDPR and AI Act exercise extraterritorial jurisdiction over projects that have users or data flows in Europe.
22 more timeline events
All Facts (20)
The UK AI Safety Institute is one of the official bodies that relatively early conducted systematic evaluations of AI model security risks on a global scale.
90%VerifiedThe EU Artificial Intelligence Act (AI Act) explicitly requires that high-risk AI systems must have human oversight and intervention capabilities.
90%Verified欧盟的《人工智能法案》（EU AI Act）于2024年正式生效，将AI系统按风险等级分类管理
90%VerifiedThe EU AI Act officially came into effect in 2024, explicitly requiring high-risk AI systems to provide explainable decision-making rationale
90%VerifiedPenalties for violations of the EU AI Act can reach up to 7% of global revenue or €35 million.
80%VerifiedThe EU AI Act has extraterritorial effect: even if a company's headquarters is not within the EU, it must comply with the regulations as long as its AI system outputs are directed at the EU market or affect EU citizens.
80%VerifiedThe EU AI Act categorizes generative AI under General Purpose AI (GPAI), requiring public disclosure of training data summaries, compliance with copyright regulations, and additional systemic risk assessments for models exceeding a specific FLOPs threshold.
80%VerifiedRegulatory frameworks such as the U.S. AI Safety Institute (AISI) and the EU AI Act are attempting to establish tiered control mechanisms targeting high-risk AI capabilities.
75%Verified欧盟《人工智能法案》（AI Act）明确要求AI生成的内容必须进行标注
75%VerifiedThe Biden administration's AI executive order required that training of models exceeding a specific compute threshold be reported to the government.
65%VerifiedBoth the EU AI Act and China's Interim Measures for the Management of Generative Artificial Intelligence Services impose compliance requirements on data processing activities by enterprises using AI tools.
65%Verified欧盟《AI法案》要求AI公司披露训练数据来源
65%VerifiedOpenAI is planning to add watermarks to text generated by ChatGPT in order to meet EU regulatory requirements for AI content provenance.
80%Unverified欧盟AI法案禁止类条款于2025年2月生效，GPAI相关条款于2025年8月生效，高风险系统条款延至2026年
90%Unverified敏感信息传入AI上下文后存在多个潜在风险面：日志留存、模型训练数据混入，以及若AI被授权网络访问时可能的外发行为
80%Unverified《人工智能法案》采用风险分级管理，将AI应用分为不可接受风险、高风险、有限风险和最小风险四个等级
75%UnverifiedThe EU's AI Act and Digital Services Act have set stricter compliance requirements for automated data collection.
60%UnverifiedThe EU AI Act applies a lower level of regulatory scrutiny to open-weight models than to closed-source models.
60%Unverified欧盟《AI法案》已将深度伪造列为高风险应用，要求生成内容必须明确标注为AI生成
60%UnverifiedThe EU AI Act has officially taken effect, requiring clear labeling of synthetic output and disclosure of bot identity to users, or otherwise facing hefty fines.
50%