衡量美国金融业的人工智能事件和影响（英）

Investment Rating - The report does not explicitly provide an investment rating for the industry, focusing instead on the risks and implications of AI in the financial sector. Core Insights - The emergence of generative and autonomous AI has significantly expanded the risk landscape for companies utilizing these advanced tools, necessitating a standardized framework to assess AI-related events and their consequences [5][6][28]. - There is a growing concern regarding the lack of transparency in reporting AI-related risks, with a notable increase in mentions of AI-related reputational risks in SEC filings, rising by 46% from 2024 [6][7]. - The report highlights the need for a tailored measurement framework for AI-related events in the financial services sector to bridge the transparency gap and establish risk models [5][28]. Summary by Sections Introduction - AI has been part of the financial technology stack for decades, but the rapid deployment of generative AI has introduced new risks that require a standardized assessment framework [5]. - Current literature and research on the business and academic implications of AI risks are scarce, emphasizing the need for robust governance measures and risk measurement frameworks [5][6]. AI Risk Awareness and Reporting - The SEC has warned companies about accurately disclosing AI-related risks, with enforcement actions against firms engaging in "AI washing" [6]. - A significant increase in the mention of AI-related risks in SEC filings has been observed, from 4% in 2020 to 43% in 2024, although these mentions often lack detail on mitigation measures [6][7]. AI Event Classification Framework - The report proposes a classification framework for AI events specific to the financial sector, which includes event types, causes, impacts, and metadata [12][13]. - Existing AI event databases are valuable for identifying and categorizing AI incidents, but they lack standardization, which limits comparative analysis [12][24]. Future Outlook - As generative and autonomous AI systems become more integrated into financial services, the industry must address the increasing risks of technical failures, governance shortcomings, and adversarial misuse [28][29]. - The proposed classification framework aims to facilitate better understanding and tracking of AI-related events, helping financial institutions build a business case for stricter regulations and establish benchmarks for AI-related risks over time [28][29].