The code which governs appraisers, the Uniform Standards of Professional Appraisal Practice, demands that if appraisers use data or mathematical models, Radke said, they must be able to explain how the model works to regulators.

“They’re scared to death of adopting new technologies because they’re afraid they’re going to have to explain it to somebody,” Radke said, of appraisers.

Some technological advancements, until recently, may have been too clunky or expensive to be within reach, said Kenon Chen, executive vice president of corporate strategy at appraisal management company Clear Capital. Now, many of those capabilities are built-in to the latest smartphones.

“The promise of photo, AI and machine learning for data being collected at the home is here, now, not a couple years from now,” said Chen. “There’s an opportunity for there to be less art and more science.”

Proponents of technology argue that such tools allow for more efficiency and standardization, while removing the potential for unconscious bias. Tech evangelists argue that relying more on data could speed turn-times for appraisals — a big plus for lenders.

The Federal Housing Finance Agency took a firm step toward addressing those concerns, when it announced this week that Fannie Mae and Freddie Mac will soon allow desktop appraisals on a permanent basis for some purchase loans.

Desktop appraisals rely on publicly-available data, such as from listings and tax assessments, and can be completed remotely. But the FHFA stopped short of allowing hybrid appraisals. Hybrid appraisals, which the agency has also called “bifurcated,” allow a third-party — an appraiser trainee, home inspector, or real estate agent — to visit the property and collect data. An appraiser in an office then uses the data to complete the appraisal.

The FHFA said last year that there are some risks to the hybrid model, “because a uniform regulatory framework does not exist at both the state and federal levels that holds non-appraisers accountable for their work on appraisals.”

In response to a question about third-party risk, Radke countered that there are deficiencies in the current model.

“We don’t necessarily get perfection from appraisers right now, although I think for the most part they’ll do a better job than an amateur,” Radke said.

The appraisal profession has found itself at the center of a reckoning on racial equity, bias and the persistent racial homeownership gap. Appraisers argue they have been unfairly targeted, when many other factors determine property values and the credit decisions that impact borrowers of color. Despite those concerns, an effort is underway at the Department of Housing and Urban Development to root out bias specifically in appraisals.

Last week, its task force on property appraisal and valuation equity hosted a closed-press industry listening session with 70 leaders from industry, advocacy, fair housing and academia organizations. The task force will make recommendations for policy actions early next year.

But questions remain over the prevalence of bias and the shape it takes in appraisals. Jillian White, head of collateral at Better Mortgage, said that Better recently started tracking bias complaints in requests for reconsideration of value. Those complaints are mostly, but not always, linked to race and ethnicity, White said.

White said she noticed bias claims spike when there are news media reports of bias, with some borrowers citing articles as a reason for their claim. Bias claims are also more prevalent in cash-out refinances, she said.

Both Fannie Mae and Freddie Mac have said they are conducting research into bias in appraisals using data from appraisal reports. Freddie Mac recently found that appraisals in Black and Latino neighborhoods are more likely to fall short of the contracted sale price.

Fannie Mae has declined to share their research, although Radke said the enterprise is “uniquely positioned to understand” whether appraisers are biased because of the loan-level and appraisal data it possesses. Neither of the GSEs have made their data publicly available. Researchers have also been unsuccessful in gaining access to the data.