Artificial Intelligence and RMBS: The SEC’s AI-Related Disclosure Requirements Trend

The U.S. Securities and Exchange Commission (SEC) has recently increased scrutiny on artificial intelligence (AI) through its comment letters, highlighting the need for transparency, governance, and risk management when AI is used in financial processes. While this trend primarily impacts companies that rely on AI for material financial decisions, it also indirectly affects due diligence providers involved in Residential Mortgage-Backed Securities (RMBS) transactions.

The Nationally Recognized Statistical Ratings Organizations (NRSROs) and RMBS investors need to understand the context of the use of AI in the mortgage manufacturing and the subsequent loan level reviews by third-party due diligence services providers (Due Diligence firms).? The recent SEC comment letters not only emphasize concerns about AI usage regarding transparency, specifically requiring clear disclosures regarding how the AI is used in credit decision-making processes, but concerns center on how AI outcomes impact understanding of financial performance and the if, and how, the AI outcomes may directly impact investor confidence. AI is increasingly being used for loan level underwriting, risk modeling, and due diligence, and the requisite RMBS disclosures must address potential risks, such as data bias, manufacturing or review inaccuracies, or systemic vulnerabilities that are inherent in AI-driven processes.

Whom Is Best Suited to Assess AI Risk?

Due diligence providers play a critical role in RMBS transactions by evaluating loan quality, ensuring regulatory compliance, and validating the accuracy of underlying asset quality and data. Their findings are typically included in third-party due diligence reports filed with the SEC under?Rule 17g-10, which mandates transparency and accountability in RMBS transactions. These providers are increasingly relying on AI-driven technologies for tasks such as loan-level risk assessments, data validation and anomaly detection, fraud detection, and with property valuation models the use machine learning algorithms. Given the SEC’s focus on AI disclosures, Due Diligence providers must evaluate how the their use of AI, and that of mortgage originators, intersects with these heightened expectations.

NRSROs and the SEC must increase scrutiny on AI-driven analyses within the manufacturing of mortgage, the RMBS structuring and the underlying collateral assessment and review processes as the AI related risk is real, and will have an impact on RMBS collateral quality.? With the increased use of AI by Due Diligence companies to assess loan quality, collateral value, or compliance risks, the NRSROs will need to ensure that the types of a AI, including the underlying models, used by the Due Diligence firms are well-documented and transparent.? Due Diligence firms should anticipate that the NRSROs, investors and regulators may have questions about the reliability, accuracy, and governance of the AI models used throughout the origination and due diligence processes, and the resulting loan level outputs, whether it be data, or credit, collateral and compliance quality. The Due Diligence firms should be prepared to provide the appropriate parties with adequate transparency and detail regarding the particular use of the AI by the originators and the Due Diligence firms, the governance and controls employed through the use of the AI, and the overall risk management process regarding the particular use of the AI throughout the manufacturing and audit of the mortgages.

Due Diligence providers must demonstrate that their AI systems, and that of the mortgage originators are free from significant biases, as well as credit, collateral and compliance risks that could undermine loan evaluations and the underlying loan level data reported to the NRSROs and investors.

RMBS issuers are ultimately responsible for the disclosures filed with the SEC. If an issuer uses AI-based Due Diligence services, and/or placing AI manufactured mortgages into the bond, issuers may require Due Diligence providers to share detailed information about the AI and its inherent risk, to appropriately align with issuer-level SEC transparency requirements. Third-party Due Diligence reports, filed as part of RMBS transactions, might need to include more detailed disclosures about how AI was used to manufacture and evaluate loan data. While Rule 17g-10 does not explicitly require AI-related disclosures, the SEC’s broader focus on AI may very well, indirectly, compel Due Diligence providers to address these factors in their reporting.

To align with the SEC’s evolving expectations, Due Diligence providers in RMBS transactions should consider how to best document AI usage, detail governance and oversight mechanisms, and work closely with issuers to ensure that AI-related disclosures in due diligence reports are consistent with broader issuer filings. This would include clearly defining how AI models are used in due diligence processes, including data inputs, algorithms, and decision-making outputs. Questions that should be addressed, include, but are not limited to… What specific tasks does AI perform? How is AI validated to ensure accuracy and reliability? And what are the limitations or potential biases of the AI models?

What is Next?

As the SEC’s focus on AI-related disclosures grows, Due Diligence providers in RMBS transactions will need to quickly adapt. By proactively enhancing transparency, improving governance, and aligning with issuers’ disclosure practices, Due Diligence providers can mitigate risks and maintain confidence in their services. Due Diligence firms are in the best position to provide the NRSROs, as well as investors, with the appropriate understanding of the risk associated with the quality of the underlying collateral so the appropriate risk ratings and investment decisions can be made, based upon the best available information.? In this uncertain, but still evolving regulatory landscape, those who embrace these changes as opportunities to build trust and reliability will stand out as leaders in the RMBS Due Diligence space as the true barometers of risk.? It is inevitable that the NRSROs and the SEC will demand such transparency to not only protect the investors, but to preserve the integrity of the securitization industry as a whole.