How AI Could Pose a Threat to the Stability of Financial Markets

Artificial intelligence (AI) is a powerful technology that has the potential to transform various sectors of the economy, including finance. However, it also comes with certain risks and challenges that need to be addressed by regulators and market participants. SEC Chair Gary Gensler recently warned about some of these issues in a virtual fireside chat hosted by Public Citizen on Jan. 17.

AI Washing and Fraud

One of the problems that Gensler highlighted was the phenomenon of AI washing, which refers to making exaggerated or misleading claims about the use or capabilities of AI models. This could happen when issuers raise money from the public or file their quarterly reports, and they are not truthful about the risks and benefits of their AI systems. Gensler said that the SEC has found that when new technologies emerge, issuers must be honest and transparent about their claims, and disclose how they manage and mitigate the risks associated with AI.

Another issue that Gensler mentioned was the possibility of fraud and manipulation using AI. He said that people could use AI to create fake or deceptive content, such as deepfakes, to mislead the public or investors. He gave an example of a fake blog post that announced his resignation in July of 2023, which he attributed to AI. He also noted that the SEC’s official Twitter account was hacked recently, and a false tweet was posted about the approval of a spot Bitcoin exchange-traded fund (ETF). He said that these incidents show that fraud is fraud, and that it is against the law to use AI or any other means to deceive the public.

AI Narrowcasting and Bias

Gensler also discussed how AI could enable a new form of narrowcasting, which means targeting specific individuals or groups based on their data and preferences. He said that this could happen when AI-enabled systems offer financial advice or services to investors, such as robo advisors or broker dealers. He warned that if these systems put their own interests ahead of the investors’ interests, or if they are not transparent about their fees and incentives, then there could be a conflict of interest or a breach of fiduciary duty.

Moreover, Gensler pointed out that AI could also introduce bias and discrimination in the financial markets, especially if the data or algorithms used by the systems are not representative or fair. He said that AI could amplify existing inequalities or create new ones, and that this could affect the access and affordability of financial products and services for certain groups of people. He said that the SEC is looking into how to ensure that AI systems are fair, inclusive, and accountable, and that they do not harm the investors or the public.

AI Centralization and Fragility

Finally, Gensler expressed his concern about the centralization and fragility of the AI market, which could have implications for the stability and resilience of the financial system. He said that he expects AI to tend toward centralization, similar to the cloud provider and search engine markets, where only a few dominant players exist. He said that he believes that there will be only a handful of large base models and data aggregators in the AI market, and that hundreds or thousands of financial actors will rely on them.

Gensler said that this could create a monoculture, where if the central nodes have it wrong, the whole system could go wrong. He said that this could lead to systemic risk and future financial crises, and that regulators do not have oversight over these central nodes. He said that regulators in the U.S. and internationally need to think about how to maintain diversity and competition in the AI market, and how to ensure that the AI systems are robust, reliable, and secure.

Gensler concluded his remarks by saying that AI is a transformative technology that could bring many benefits to the financial markets, but also many challenges and risks that need to be addressed. He said that the SEC is working to protect the investors and the public from the potential harms of AI, and to promote innovation and competition in the AI market. He said that he welcomes the input and collaboration of all stakeholders, including the industry, the academia, the civil society, and the public, in this endeavor.