Cotality’s AI in Housing 2026 report, released in April, starts from a number that sounds like unqualified adoption: 75% of buyers now assume artificial intelligence plays some role in their home purchase. Read past the headline, though, and the report describes something closer to acceptance without confidence, buyers who expect AI to be everywhere in the process and trust almost none of it to make a decision on their own.
Where buyers already assume AI is involved
The 75% figure breaks down by category, and it isn’t evenly distributed. Buyers most commonly assume AI is used by property websites (86%), insurers (82%), lenders (80%), real estate agents (80%), and brokers (79%). Buyers have stopped asking whether AI touches their transaction. They now assume it touches nearly every part of it, from the listing they clicked on to the rate quote they received.
Confidence and trust are moving in opposite directions
That assumption hasn’t translated into comfort. Cotality found that overall buyer confidence in navigating the homebuying process fell from 83% in 2025 to 72% in 2026, a drop the company ties directly to the speed of AI adoption outpacing buyers’ ability to evaluate what the tools are actually doing. Trust in AI specifically to help find or evaluate a home fell even further, from 30% to 16% year over year, nearly cutting the figure in half in twelve months.
Usage is climbing. Trust is collapsing. That contradiction is the report’s central tension. Cotality’s own data science lead framed it as buyers wanting the speed AI offers without giving up certainty about what they’re actually agreeing to.
The generational split
Younger buyers are both the heaviest users of AI tools and the most willing to credit them with real value. Half of Gen Z buyers say AI increases their confidence in the homebuying process, compared with 40% of millennials, 33% of Gen X, and just 21% of baby boomers. That same younger cohort also drives the report’s next finding, since millennials and Gen Z are simultaneously the most frequent AI users and the most likely to pay for a human to double-check what the AI produced.
Buyers want disclosure, not just accuracy
A related demand sits underneath the trust numbers: buyers want to know when AI is involved, not just whether it got the answer right. Sixty-eight percent said clear labeling of AI-generated property listings and mortgage recommendations is important or essential, and 37% said that labeling should be mandatory, a figure that climbs to 61% among baby boomers specifically. The same unease shows up in how buyers view the data behind the AI itself: 64% worried that AI tools are recycling unverified information across multiple sites rather than pulling from validated, first-party sources.
Tolerance for AI getting something wrong follows the same generational split as trust does. Twenty-two percent of Gen Z buyers and 19% of millennials say they’re tolerant of AI errors, compared with just 11% of Gen X and 9% of baby boomers, the same generations least likely to say AI increases their confidence in the first place.
Buyers want AI’s speed and a human’s judgment
Forty-four percent of buyers surveyed said they’d pay an additional fee for a human expert to verify an AI-generated housing decision, and the tasks they want a person handling aren’t evenly weighted either. Sixty-six percent said they’d rather rely on a human than AI for legal assistance, up from 54% just a year earlier. Fifty-six percent said the same about assessing natural disaster risk, and a majority said they’d prefer a person securing their mortgage rather than software doing it alone.
That preference maps closely onto where a buyer’s exposure runs highest: the legal and financial steps where a wrong answer doesn’t surface until it’s expensive to fix, exactly the post-purchase gap in first-time buyer education that AI usage alone hasn’t closed. Cotality describes this as buyers asking for safeguards before letting AI touch the largest financial decision most of them will ever make.
None of this suggests AI is losing ground in homebuying. Usage keeps climbing across every category Cotality measured. What’s climbing alongside it is a specific, measurable demand for a human being to confirm the parts of the process that carry the most consequence if AI gets them wrong, a demand that grew sharply in exactly the twelve months trust in AI fell the furthest.
