The U.S. Department of the Treasury on Thursday initiated an inquiry into the integration of artificial intelligence in financial services. This move coincides with the Office of the Comptroller of the Currency's acting chief's warning that a potential "nightmare" scenario involving AI on Wall Street is "uncomfortably plausible."

The Treasury issued a comprehensive request for information to understand how AI is currently being utilized by financial institutions, its potential future applications, and the associated risks for banks, businesses, regulators, consumers, and other market participants. This inquiry, open for a 60-day comment period, is part of the Biden administration's broader efforts to regulate AI and its fast-evolving applications across various industries.

Treasury Secretary Janet Yellen emphasized the urgency of this initiative, noting that AI's rapid development presents significant opportunities and risks for the financial sector. Yellen highlighted that the Financial Stability Oversight Council (FSOC), which she chairs, is closely monitoring AI's impact on financial stability and is working on building supervisory capacity to better understand these risks.

Federal financial regulators have flagged multiple concerns regarding AI, such as the potential for perpetuating algorithmic bias and facilitating fraud through deepfakes. Despite these risks, there is also recognition of AI's potential benefits, including improved detection of illicit activities, enhanced back-office efficiency, and more inclusive underwriting processes.

Acting Comptroller of the Currency Michael Hsu cautioned that AI's ability to learn sets it apart, making it easier to disclaim responsibility for adverse outcomes. He suggested adopting a "shared responsibility model" from cloud computing, where safety and security responsibilities are clearly divided between financial firms and AI service providers.

Hsu argued that past financial innovations, like derivatives and cryptocurrencies, went off the rails due to rapid adoption without sufficient controls. He warned that AI might follow a similar trajectory if not properly managed. He urged banks to establish checkpoints to ensure responsible AI innovation and prevent runaway growth.

Hsu also highlighted the growing risks of AI being used for fraudulent purposes and cyberattacks. He expressed concern over AI-generated disinformation and its potential to amplify financial system vulnerabilities. Hsu referenced a hypothetical scenario where an AI-powered trading agent could manipulate markets by spreading false information to maximize returns, a situation he finds disturbingly plausible.

Overall, the Treasury's inquiry aims to facilitate responsible AI use in financial services while addressing potential threats to financial stability, ensuring a balanced approach to innovation and regulation.