人工智能（AI）：国际金融监管初探

Industry Investment Rating - The report does not explicitly provide an investment rating for the AI industry [1][2][3] Core Viewpoints - AI has been widely applied across various fields, with significant potential demonstrated by the launch of ChatGPT in 2022 [2] - AI technology poses risks such as systemic discrimination, privacy violations, and information bubbles, prompting global regulatory efforts [2] - Different regions, including the EU, UK, US, and China, have adopted varying regulatory approaches based on their technological strengths and resources [2] - The EU has taken a leading role in AI regulation, introducing ethical guidelines, comprehensive legislation, and international conventions [4][16] - The UK focuses on innovation-friendly regulation, avoiding hasty legislation and favoring soft frameworks and guidelines [24][25] - The US emphasizes industry self-regulation and collaboration between government and private sectors, with state-level legislation being more active than federal [32][33] - China has initiated early planning and governance activities in AI, with a focus on ethical norms and international cooperation [42] Regulatory Overview - Globally, 18% of AI regulations are binding, with the majority being soft guidelines, principles, and discussion documents [4] - The EU has a comprehensive AI regulatory system, covering planning, principles, ethics, frameworks, and rules [4][16] - The UK adopts a context-specific, risk-based, and innovation-supportive regulatory approach [25][27] - The US has shifted from voluntary guidelines to administrative actions, with state-level legislation being more active [32][33] - China has a limited number of binding AI regulations, focusing on planning, governance principles, and ethical norms [42] Key Regulatory Documents - EU: - Ethics Guidelines for Trustworthy AI (2019) outlines a lifecycle framework for trustworthy AI [16] - EU AI Act (2024) is the first comprehensive global AI regulation, introducing a risk-based approach [20] - Framework Convention on AI, Human Rights, Democracy, and the Rule of Law (2024) is the first legally binding international treaty on responsible AI use [22] - UK: - National AI Strategy (2021) outlines a 10-year plan to make the UK an AI superpower [25] - Pro-innovation Approach to AI Regulation (2023) proposes a principles-based, flexible regulatory framework [27] - US: - Blueprint for an AI Bill of Rights (2022) provides principles for protecting public rights in AI systems [33] - AI Risk Management Framework (2023) offers voluntary guidelines for AI risk management [34] - Executive Order on Safe, Secure, and Trustworthy AI (2023) emphasizes transparency and accountability in AI development [37] - China: - Next Generation AI Development Plan (2017) sets the foundation for AI governance and ethical norms [42] - Global AI Governance Initiative (2023) and Shanghai Declaration on AI Global Governance (2024) highlight China's role in international AI governance [42] Financial Sector AI Regulation - Financial regulators globally are gradually introducing AI-specific regulations, with the EU leading in integrating AI into financial oversight [13][23] - The UK financial regulators leverage existing frameworks to manage AI risks, avoiding the creation of new, isolated regulatory structures [31] - The US financial sector has limited binding AI regulations, with ongoing discussions and expected future rulemaking [39] - China's financial regulators are expected to align with global AI governance trends, with more regulations anticipated in the future [42]