Mortgage Industry Trends: AI & ML Insights from Fannie Mae Report

TL;DR:

The 2023 Fannie Mae Mortgage Lender Sentiment Survey provides a comprehensive analysis of how artificial intelligence and machine learning are being integrated into the housing finance industry. By comparing current data to a 2018 study, the report reveals that while general familiarity with these technologies has remained steady at approximately 65%, the primary motivation for adoption has shifted toward improving operational efficiency. Modern lenders are specifically interested in utilizing automated tools for compliance management, underwriting data verification, and property appraisals. Despite this interest, significant barriers to widespread implementation persist, including the complexity of integrating new software with legacy infrastructure, high costs, and a lack of established success stories. Additionally, the survey highlights emerging anxieties regarding data security and the potential for misinformation generated by automated systems. Ultimately, while full-scale deployment remains limited, most institutions expect to increase their investigative efforts or trial usage within the next two years.

AI/ML Familiarity and Adoption Trends (2023 vs. 2018)

While familiarity remains high, the pace of broad deployment has slowed, with more lenders shifting toward limited trials.

Metric	Q3 2023	Q3 2018
Familiar with AI/ML technology	65%	63%
Deployed or trial users of AI/ML	30%	27%
Have not yet looked into AI/ML	29%	37%
Expect to broadly roll out/start trials in 2 years	55%	58%
Very Familiar with AI/ML	16%	16%
Current Users (fully incorporated tools)	7%	14%
Trial Users (limited or trial basis)	22%	13%

Primary Objectives and Motivations

There has been a massive shift in motivation toward operational efficiency and away from consumer-facing improvements.

• Improving Operational Efficiency: 73% of lenders cited this as their top objective in 2023, compared to only 42% in 2018.

• Enhancing Consumer Experience: This objective plummeted from 41% in 2018 to just 7% in 2023.

• Reducing Human Error: Remained steady at 9% in both survey years.

• Better Control of Risks: 5% in 2023 vs. 7% in 2018.

Top Challenges to AI/ML Adoption

Lenders who have not yet used AI/ML applications cited the following as their biggest combined challenges (sum of first and second biggest):

1. Integration Complexity: 48% (e.g., legacy systems not built for AI).

2. Lack of Proven Success Record: 35%.

3. Costs Too High: 24%.

4. Security/Privacy Concerns: 22% (significantly increased from 10% in 2018).

5. Potential for Bias/Discrimination: 15%.

Risk Assessment and Application Appeal

Lenders view misinformation as the most significant threat, while they are most interested in AI for compliance and fraud detection.

Perceived Biggest Risks (Select One):

• Misinformation: 26%.

• Cybersecurity: 18%.

• Bias and Discrimination: 16%.

• Privacy Concerns (Consumer Data): 15%.

• Transparency and Accountability: 14%.

• Job Displacement: Only 5% view this as the biggest risk.

Most vs. Least Appealing AI Applications (Combined top 2):

• Most Appealing: Compliance Review (50%) and Anomaly Detection/Fraud Automation (39%).

• Least Appealing: Virtual Assistants/Chatbots (35% cited as least appealing) and Borrower Prepay Assessment (21%).

Sub-group Statistics: Lender Size and Type

Adoption and familiarity vary significantly based on the size of the institution and its business model.

• Familiarity: 63% of larger/mid-sized institutions expect to use AI in two years, compared to 40% of smaller institutions.

• Current Non-Usage: 50% of smaller institutions have not used or explored AI, compared to 18% of larger and 20% of mid-sized institutions.

• Institution Type: Credit Unions (28%) are much more likely than depository institutions (7%) to be “very familiar” with AI/ML.

• Integration Hurdles: Larger institutions (55%) are more likely than smaller ones (37%) to cite integration complexity as a primary barrier.

Economic and Housing Sentiment (Q3 2023)

The sources also provide context on the broader market in which these lenders operate.

• U.S. Economy Outlook: 54% of lenders believe the economy is on the wrong track, while 39% believe it is on the right track.

• Ease of Getting a Mortgage: 75% of lenders believe it is currently difficult for consumers to get a mortgage, up from 53% in Q3 2022.

• Home Price Predictions (Next 12 Months):

→ Go Up: 32%.

→ Stay the Same: 47%.

→ Go Down: 20%.

• Price Increase/Decrease Magnitude: Lenders who expect prices to rise anticipate a mean increase of 4.29%. Those who expect a drop anticipate a mean decrease of 6.38%.

Citation

Data and insights referenced in this article are sourced from the 2023 Fannie Mae Mortgage Lender Sentiment Survey, which provides an in-depth analysis of artificial intelligence and machine learning adoption trends in the mortgage industry. The full report is available on Fannie Mae’s official website: Mortgage Lender Sentiment Survey® Special Topics Report