Is the Mortgage Industry Prepared for the “Dark Side” of AI?

By Martin S. Frenkel

As Darth Vader once said, “You don’t know the power of the dark side.” Or, for the purposes of this article, “you don’t know the potential impact of the ‘dark side’.”

AI is all around us – even if it is not significantly impacting our daily life (yet). There is a constant diet of AI articles in the media and the predictions that come along with them. AI will reduce costs, making businesses more profitable, or improve the consumer experience. AI will cure cancer or solve the puzzle of nuclear fusion. AI will allow us to perform innumerable tasks better, faster, and with fewer errors, providing us with greater leisure time not realized since the Industrial Revolution. All, some, or none of these predictions may ultimately prove out. The fact is that no one really seems to know when, what, how, or to what degree the impact of AI will be felt.

However, one increasingly present theme is the probability of a rapid and massive disruption to the workforce, particularly among white collar workers.¹ According to certain sources, and depending upon the definition one applies to white collar worker (often defined as jobs in management, the professions, and related fields), approximately 55-60% of the workforce is white collar.² 

What if these dire predictions of rapid and immense job losses for white collar workers come true? How will this impact the mortgage industry? Will this impact be the same, or differ from, normal recessions?

In evaluating the answers to these questions, let’s start with one fact which is axiomatic to the mortgage industry: employment is a critical factor in the mortgage underwriting process because it is essential to evaluating a borrower’s ability to repay a loan.³ On the servicing side, employment and default rates are also intimately linked, with defaults rising in times of increased unemployment.⁴ Simply put, a borrower’s employment is key to their ability to get a loan and to their ability to continue to repay that loan over time.

If AI causes a rapid and vast disturbance within the labor force, there is a risk that the mortgage industry will be hit with the proverbial double-whammy: a significant reduction in qualified borrowers for new loans and spiking defaults on existing loans. Is this really likely to happen? The truth appears to be that no one really knows.

Perhaps history can provide some basis to evaluate the issue. The mortgage industry has weathered recessions repeatedly and knows how to scale down in recessionary periods and up in recovery periods. Since World War II, there have been numerous recessions, with the average period being 11.1 months before a recovery began.⁵ Surely, the industry is well-versed in navigating this type of economic disruption.

However, what if the disruption caused by AI is not like your average recession? What lessons can be learned from history? COVID caused unemployment to spike at 14.8%, but this high unemployment rate lasted only approximately a year.⁶ During the Great Recession, unemployment peaked at 10%, but employment largely recovered over the following two-year period.⁷ 

What about the Industrial Revolution as a historical parallel? The Industrial Revolution is known to have caused substantial job losses where manual jobs were replaced by machines. However, as disruptive as this period was, it created many more new types of jobs and took place gradually over a period of 150 years, allowing for slow adjustment within the workforce. Similarly, large numbers of jobs were lost during the first computer revolution of the 1970s and 1980s, where hundreds of thousands of clerical and manufacturing jobs were lost. Again, though, this transition took place over a couple of decades, allowing workers to retrain and adjust gradually.  

Perhaps none of these events will prove to be the historical parallel to the impact of AI on employment. The Great Depression is yet another terrifying potential measuring stick. Although not caused by technological innovation, the Great Depression began suddenly in 1929, resulting in enormous unemployment rates. At its height in 1933, 24.9% of the total workforce was unemployed.⁸  Furthermore, the Great Depression lasted for twelve years.⁹  Similarly, some estimates predict that AI will result in rapid job losses, replacing approximately 50% of all existing jobs by 2045.¹⁰  The loss of jobs for some may be permanent.

If the constant buzz surrounding AI is any indication, we are approaching a historical inflection point. Just as no one could predict the true impact of these past historical events, we cannot accurately predict the impact of AI. Will AI usher in a new era of prosperity and be positively transformative for employment, have a hugely disruptive effect on employment more akin to the Great Depression, or be a “nothing burger” like Y2K?  

Right now, this article is merely a thought exercise. But, soon we as a society and, more specifically, as the mortgage industry, will face an altered employment reality – whether it be minor or massive. These are truly fascinating times in which to live.