On August 6, 2026, the conversation around artificial intelligence in mortgage lending changed.
That is the effective date of Fannie Mae's Lender Letter LL-2026-04, establishing governance expectations for lenders using artificial intelligence and machine learning in connection with loans sold to Fannie Mae. Freddie Mac has already implemented similar requirements within its Seller/Servicer Guide. Together, the Government-Sponsored Enterprises have sent a clear message to the mortgage industry: AI governance is no longer optional.
For many lenders, however, the question isn't whether they agree with the need for governance. The question is much more practical:
Where Do We Begin?
That question has become one of the most common themes I hear from mortgage executives across the country. Nearly every lender is using AI today in some form. Sometimes intentionally. Sometimes without realizing it. Artificial intelligence now exists inside loan origination systems, document recognition software, fraud detection platforms, servicing applications, pricing engines, CRM systems, quality control tools, and employee copilots. Vendors continue introducing new AI capabilities through routine software updates, often without requiring lenders to purchase a separate AI solution. The reality is simple. You cannot govern what you haven't identified.
Existing Laws Already Apply
One misconception continues to surface throughout the industry: that lenders should wait for new AI regulations before building governance programs. That is not what the market is telling us. Earlier this year, the Mortgage Bankers Association published Examining AI-Powered Mortgage Through the Lens of Federal Law, concluding that existing mortgage laws, including ECOA, the Fair Housing Act, FCRA, GLBA, HMDA, RESPA, Regulation Z and UDAAP, already govern AI-enabled mortgage activities. AI doesn't replace existing regulatory obligations. It operates within them.
The GSEs have now reinforced that expectation. Freddie Mac's Seller/Servicer Guide Section 1302.8 requires seller-servicers using AI or machine learning to maintain documented governance, establish executive accountability, implement risk management practices, and provide information regarding AI safeguards when requested. Fannie Mae's Lender Letter LL-2026-04, effective August 6, similarly establishes governance expectations surrounding the use of AI and machine learning in both origination and servicing. Neither organization is attempting to slow innovation. Instead, both are asking lenders to demonstrate responsible oversight.
Governance Is Becoming A Competitive Advantage
Too often, governance is viewed as something that happens after innovation. The opposite is true. Organizations that establish governance early gain the confidence to deploy AI more broadly, onboard new technologies faster, satisfy investors and regulators more efficiently, and make better-informed purchasing decisions. Governance should not become the brake. It should become the accelerator.
That philosophy guided the development of FRAME (Framework for Responsible AI in Mortgage Ecosystems), developed by MISMO in collaboration with the mortgage industry's AI Community of Practice and subject matter experts. FRAME was never intended to create new regulatory obligations. Its purpose is to help mortgage companies operationalize governance using practical tools designed specifically for residential mortgage lending while aligning with broader frameworks such as the NIST AI Risk Management Framework.
Rather than asking lenders to begin with a lengthy policy manual, FRAME starts with six practical steps:
Find. Inventory. Tier. Assess. Monitor. Update.
Identify where AI exists. Document each use case. Assign ownership. Evaluate risk. Monitor performance. Continuously improve governance over time. It sounds straightforward because it is. The challenge has never been understanding why governance matters. The challenge has been knowing how to build it.
From Compliance Requirement To Implementation
Recognizing the industry's need for practical implementation guidance, MISMO will dedicate the opening day of the upcoming MISMO Summit to AI governance.
On August 24, mortgage professionals from across the industry will gather in Reston, Virginia, for an immersive FRAME AI Governance Workshop focused entirely on implementation. This is intentionally designed as a working session.
Participants will build AI inventories, identify AI use cases throughout the mortgage lifecycle, evaluate risk, assign ownership, and begin constructing governance programs they can immediately take back to their organizations. The goal is not simply to understand FRAME. It is to begin using it.
Whether you're a lender, servicer, technology provider, compliance executive, risk officer, or operations leader, the objective is the same: leave with tangible progress instead of another notebook full of ideas.
Building An Industry Standard
AI governance is ultimately about people as much as technology. Mortgage organizations need professionals who understand how to identify AI systems, assess risk, document oversight, and maintain governance programs over time. They also need greater confidence that the vendors supporting them have established responsible AI practices. That is why MISMO will announce two new certification programs during the August Summit.
The first is an AI Governance Practitioner Certification, designed for organizations implementing mortgage AI governance programs using FRAME. The second is a Vendor AI Governance Certification, intended to establish a common industry baseline for mortgage technology providers and service partners. These certifications are not intended to replace a lender's own due diligence. Rather, they are designed to strengthen it by creating more consistent governance practices, documentation and evidence across the mortgage ecosystem.
August 6 Is Only The Beginning
The effective date of Fannie Mae's governance framework should not create panic. It should create momentum. No lender is expected to build a mature AI governance program overnight. But every lender should know where AI exists within its organization, understand which systems create the greatest operational or regulatory risk, assign accountability, and begin documenting governance decisions. Mortgage lending has always depended upon trust. Artificial intelligence doesn't change that. It simply raises the standard for how trust is demonstrated.
August 6 marks an important milestone. August 24 provides the opportunity to begin building the governance framework that will support the next generation of mortgage lending.
To learn more about the AI Governance Workshop and register for the MISMO Summit, visit:
https://www.mismo.org/events-education/mismo-summit
