MFA outlines approach to AI oversight in capital markets

Washington, D.C. - MFA highlighted to Congress how existing securities laws address the risks and opportunities associated with artificial intelligence (AI) in a letter sent Friday to the House Financial Services Committee. Alternative asset managers are using AI across investment, compliance, cybersecurity, and operational functions, creating opportunities to improve efficiency, strengthen controls, and better serve institutional investors.

"Alternative investment advisers are deploying AI to improve portfolio management, risk controls, and investment research on behalf of the pensions, foundations, and endowments they serve," said Bryan Corbett, MFA President and CEO. "It is critical to maintain the flexible, principles-based approach to regulation that has long fostered innovation as AI continues to unlock new opportunities throughout the asset management industry. Existing SEC regulation and investment advisers' fiduciary duties provide a comprehensive framework for governing the development and use of these important tools."

The Investment Advisers Act of 1940 and Securities and Exchange Commission (SEC) rules hold investment advisers accountable for how they develop, deploy, and oversee AI, as they do for other technologies. Advisers must meet fiduciary, compliance, disclosure, recordkeeping, trading, and supervisory obligations regardless of whether they use traditional software, quantitative models, or AI-enabled tools. These principles-based requirements allow advisers to tailor governance and risk controls to different applications while avoiding prescriptive rules that can quickly become outdated. Creating a separate regulatory regime for AI would be a departure from how securities laws have long accommodated technological change.

The letter also urges policymakers to account for differences among financial institutions and AI use cases. Alternative asset managers operate differently from large, systemically important banks, including through matched asset and liability structures, comparatively lower leverage, and a sophisticated institutional investor base. Policymakers should avoid one-size-fits-all requirements that ignore these differences and the varying risks of different AI applications.

Read the full letter here.

###