Acting Comptroller of the Currency Calls for Shared Responsibility in AI Adoption

Introduction

Acting Comptroller of the Currency Michael Hsu emphasized the need for artificial intelligence (AI) providers and end-users, particularly banks, to establish a framework of shared responsibility for errors arising from the adoption of AI models. Hsu addressed this topic during his speech at the 2024 Conference on Artificial Intelligence and Financial Stability, hosted by the Financial Stability Oversight Council and the Brookings Institution.

Shared Responsibility Framework for AI Adoption

Hsu suggested that a similar shared responsibility model, as implemented in cloud computing, could be developed for AI adoption. This model allocates operations, maintenance, and security responsibilities to customers and cloud service providers based on the selected service. Hsu urged AI providers and customers to collaborate in creating a shared-responsibility framework, citing that they should be accountable for any errors resulting from reliance on AI models.

The Role of the U.S. Artificial Intelligence Safety Institute

Hsu proposed that the U.S. Artificial Intelligence Safety Institute, under the Department of Commerce’s National Institute of Standards and Technology, could be tasked with defining the specifics of the shared-responsibility framework. This organization, established in 2023, focuses on enhancing the safety and reliability of AI technologies.

Risks and Benefits of AI: Treasury Department’s Request for Information

Notably, Hsu’s remarks coincided with the Treasury Department’s request for information from the public concerning the risks and potential benefits of AI. The agency seeks to understand how to leverage the benefits of AI while mitigating associated risks, aiming to streamline processes and enhance economic inclusion.

Prudent Approach to AI Adoption in the Financial Sector

Hsu drew parallels between AI’s current adoption in the financial industry and the trajectory of electronic trading two decades ago. He emphasized the importance of establishing safeguards or “gates” between each stage of AI implementation. These gates enable firms to pause and assess the role of AI before proceeding further. Hsu urged banks to implement appropriate controls and ensure accountability based on their risk exposures and business activities.

AI and Financial Stability Risks

Hsu highlighted the potential financial stability risks associated with AI. He emphasized how AI may offer bad actors the ability to execute sophisticated attacks and scams, urging caution in AI adoption. While firms desire enhanced operational efficiency and profitability, Hsu acknowledged that consumers prioritize safety and security over minor inefficiencies.

Liabilities and Fairness of AI Decision-Making

Hsu raised concerns regarding the expanding liabilities resulting from the use of AI. He referenced a case in Canada where incorrect information provided by an AI-driven chatbot led to the airline being held liable for the error. Hsu also remarked on the challenges of holding AI systems accountable due to their complex and opaque nature. Additionally, he highlighted the potential scrutiny surrounding AI-assisted credit underwriting, as consumers denied credit based on AI decisions may question the fairness of such outcomes.

Conclusion

Hsu stressed the significance of trust in both the banking industry and the adoption of AI. He asserted that the establishment of a shared-responsibility framework for AI adoption would be crucial in mitigating risks, enhancing accountability, and ensuring fairness. As the financial sector continues to explore the potential of AI, Hsu’s call for shared responsibility aims to foster a responsible and secure AI ecosystem.

Question and Answer

1. What is Acting Comptroller of the Currency Michael Hsu’s stance on AI adoption?

Acting Comptroller of the Currency Michael Hsu believes that artificial intelligence (AI) providers and end-users, such as banks, should share responsibility for errors stemming from AI adoption.

2. Which framework does Hsu propose for AI adoption?

Hsu suggests the development of a shared-responsibility framework, similar to the model used in cloud computing, where operations, maintenance, and security responsibilities are allocated between AI providers and customers.

3. Which agency could take up the task of defining the shared-responsibility framework?

Hsu recommends that the U.S. Artificial Intelligence Safety Institute, a department of the National Institute of Standards and Technology, lead the efforts in defining the specifics of the shared-responsibility framework.

4. What is the focus of the Treasury Department’s request for information?

The Treasury Department is seeking public input to better understand the risks and potential benefits of AI. This information will help guide efforts to harness AI’s potential while mitigating associated risks.

5. How does Hsu address the potential liabilities and fairness of AI decision-making?

Hsu highlights the expanding liabilities resulting from AI use, citing cases where AI systems have provided incorrect information. He also recognizes the need to ensure fairness in AI-assisted credit underwriting, as consumers denied credit based on AI decisions may question the fairness of such outcomes.