EU structure

The European Commission unveiled a new proposal for an EU regulatory framework on artificial intelligence (AI) in April 2021. The draft AI act is the first ever attempt to enact a horizontal regulation of AI. The proposed legal framework focuses on the specific utilisation of AI systems and associated risks. The Commission proposes to establish a technology-neutral definition of AI systems in EU law and to lay down a classification for AI systems with different requirements and obligations tailored on a 'risk-based approach'. Some AI systems presenting 'unacceptable' risks would be prohibited. A wide range of 'high-risk' AI systems would be authorised, but subject to a set of requirements and obligations to gain access to the EU market. Those AI systems presenting only 'limited risk' would be subject to very light transparency obligations. While generally supporting the Commission's proposal, stakeholders and experts call for a number of amendments, including revising the definition of AI systems, broadening the list of prohibited AI systems, strengthening enforcement and redress mechanisms and ensuring proper democratic oversight of the design and implementation of EU AI regulation. First edition. The 'EU Legislation in Progress' briefings are updated at key stages throughout the legislative procedure.

我们认为,与GDPR相同,这部欧盟《AI法案》的提案以保护人的基本权利为出发点,以AI系统对公民基本权利的威胁程度对系统进行划分,对不同等级的AI系统提出了不同的合规要求,并施以不同强度的监管。若将来顺利出台,势必会成为欧盟人权法律体系的又一重要法案。我们可以看到,《AI法案》的关注重点与GDPR高度相似,例如人的年龄、种族、身体状况等可能造成人在社会中遭受歧视的因素,都被视作划分AI系统风险的依据。但从另一方面来看,《AI法案》对于某些AI系统和技术的监管可能相对严格,比如完全禁止某些AI技术的使用,有可能在全球各国紧锣密鼓进行AI战略部署、大力支持AI产业布局的大背景下,一定程度上阻滞欧洲AI技术的落地和AI产业的发展,从而对欧盟经济社会的发展产生一定的负面影响。但《AI法案》以**“安全”和“非歧视”**为原则,对AI进行风险分级的立法手法,也许会给世界其他各国AI立法提供启发。《AI法案》目前尚仍处于欧盟委员会的提案阶段,但EDPB和EDPS在6月21日的联合意见中已对该法案表现出了较为肯定的态度。尽管如此,根据欧盟立法程序,《AI法案》还需经过欧洲议会和欧盟理事会审议并修改,到最后颁布生效还需要两至三年的时间。
美国对人工智能的监管仍处于“碎片化”状态。特朗普政府时期直接将人工智能监管责任下放给不同监管机构。拜登政府或许会更重视此问题,以避免潜在风险,但基本上还是会采取分散型监管方式。联邦层面仍然缺乏统一立法。类似欧盟这种基于风险分类的整体型监管模式不太可能被采纳。