Бесплатная игра в игровом портале вулкан

Automation haunts many discussions about the future of work, employment, and the economy. But technological advances may soon hit homes in an unexpected way: could real estate appraisers be replaced by robots?

That’s the conclusion of a recent article in Bloomberg, which discusses how advances in big data and computing are helping automate this knowledge-based job, perhaps a harbinger of how advances in machine learning mean an ever-widening circle of professions are at risk.

The future of the profession has become a topic due to a recent decision by Fannie Mae and Freddie Mac, two institutions that facilitate the flow of funding for home loans nationwide. In the past, both of these entities have occasionally allowed appraisal waivers when evaluating low-cost loans. But recently, they’ve changed their stance, starting up a program earlier this year that would waive the new appraisal requirement for homes where the loan-to-value ratio is low. Instead, they’ll accept any appraisals on file from the last five years. In June, Freddie Mac said it would start accepting automated valuations for some refinancing loans.

More from Curbed:

Mortgage interest tax break has ‘no effect’ on homeownership, says study
Can a grassroots movement fix urban housing shortages?
Multibillion-dollar buying spree by Chinese investors shows signs of slowing

This decision will reduce the number of appraisals being requested, says Appraisal Institute President Jim Amorin, and implicitly suggests that a model with less human participation is just as good.

“There’s no replacement for an appraisal in most cases,” Amorin tells Curbed. “Many of the computer models use public information that hasn’t been verified. Even Zillow will tell you it’s an estimate, not an appraisal.”

These policy shifts come during an unfortunate time for a profession watching its workforce slowly shrink. Today, there are currently 78,000 licensed appraisers in the U.S., says Amorin, whose organization represents roughly 20,000 of them. That is a steep drop from the 120,000 that performed the job five years ago.

Part of the decline is due to appraisers requiring extensive training and apprenticeships to become licensed, and part is due to diminishing fees, a result of the growth of appraisal management companies that work with lenders and take a portion of the final fee. The median age of an appraiser is roughly 52-55, says Amorin, suggesting the workforce is aging, retiring, and not being replenished.

“If numbers continue the way they are, there may not be enough appraisers to meet the needs of the marketplace,” says Amorin. “We worry about how the automated models will serve the needs of consumers.”

At the same time, the technology now being cast as competition for appraisers is getting better and better. According to a Zillow engineer, the company’s Zestimate tool uses algorothms, machine learning, public records, MLS data, and information from brokers and users to create increasingly accurate value estimates. The models are continuously being trained on a daily basis to become more sophisticated; some are examining external imagesto better determine the “curb appeal” of a home Zillow even launched a $1 million Zillow Prize in May, similar to the Netflix Prize, to entice data scientists and researchers to improve the company’s algorithm and devise a more accurate method of estimating home values.

Zillow representatives noted multiple times they believe appraisals are valuable and in no way seek to replace the need for an appraisal.

Amorin believes that automated appraisals still focus too heavily on public data and often miss the little details and true picture of a property that creates an accurate value estimate. He believes you get what you pay for with automated models, and the work of an impartial appraiser is key to a functioning, transparent market.

But that doesn’t make him anti-technology. Amorin believes the future is in a marriage of man and machine, where humans and computer models combine for more accurate estimates. Appraisers get data that saves them time, while their estimates can be fed back into the algortihms and machine learning systems to make the estimates more accurate.

“If appraisers believe they can move forward doing what they’ve always done, they’ll go the way of the one-hour photo shop,” he says. “We have to adapt.”