Have you seen a photo recently and wondered if it is real or, perhaps, that you’ve seen something very similar before? Maybe you’re hunting for a home online and a photo of the “perfect” place appears on your screen — then you think that, maybe, it’s “too” perfect.
When you check out that house in person, you realize it was more than intuition: The house is different than the photos.
It’s a sign of the modern high-tech world of home selling — AI-generated photographs, according to Yahoo! Finance, which said that those AI-generated photos have “uncanny similarities,” such as suspiciously smooth exteriors, artificially green grass, distorted lines surrounding doors or windows, for example.
Yahoo attributed those characteristics as signs of heavy-handed AI-edited real estate listing photos.
Today, Yahoo said, AI editing tools can make it easier to virtually stage homes — and go even further by cleaning up the appearance of aging exteriors or sometimes unrealistically envisioning what a fixer-upper might look like after a pricy gut renovation. The tools are so new that only a patchwork of rules and regulations — often put in place by local real estate boards or multiple listing services — exist to govern their use, Yahoo said.
AI Photos Can Alter Reality
Megan Kolstad, a real estate agent with The Hive in St. Paul, Minnesota, told Yahoo that she recently toured a home in the city with a client, only to realize that AI images in the listing had added a nonexistent window in a bedroom.
“That felt misleading,” Kolstad said. “When we got to the property, it was just a lot of comparing the photos we were seeing online to what we were experiencing in real life, and I don’t feel like that’s the best use of our time.”
While Kolstad said she isn’t opposed to all uses of AI in listing photos, she said she thinks virtual staging can help buyers imagine what an empty space looks like furnished. Kolstad said she’s also encountered instances where that goes too far, such as when an eight-person sectional backs up to a medium-size window, suggesting some impossible furniture Tetris.
“Realistically, a couch of that size isn’t going to fit in that space,” she said.
Sonia Rodriguez, a Realtor with the Redux Group in Northern Virginia, told Yahoo that she found herself duped by virtual staging when she stopped to preview a listing for a client. She was prepared to show up to an empty home, but instead found a mess — the existing occupant’s belongings were there and the house had markedly different lighting. Rodriguez said she didn’t bring her client back for a tour.
“I went to go preview it, and I was like, ‘Oh, this is a no,” Rodriguez said, noting she was glad to have a chance to look at the listing before her client.
“We would have both walked into a surprise,” she said.
Yahoo reported that the National Association of Realtors says it “supports responsible AI use in real estate.” In 2022, NAR said there are ethical ways to use AI-generated photos but cautioned its members to consider the Realtor Code of Ethics, which tells them to “refrain from exaggerating, misrepresenting, or concealing pertinent facts related to a property or to a transaction.”
Real-looking AI-generated photos present a dilemma for more than just real estate listings.
Learning to Recognize Fake Photos
For example, researchers from the University of Reading, the University of Greenwich, the University of Leeds, and the University of Lincoln in the U.K. said they tested 664 participants on their ability to distinguish real human faces from images generated by AI software known as StyleGAN3.
In the study published last month in the journal Royal Society Open Science, the researchers said that AI-generated faces have reached a level of realism that defeats even “super recognizers,” a small group of people with exceptionally strong facial recognition skills.
According to the scientists, super recognizers performed no better than chance when attempting to identify which faces weren’t real. They suggest a few minutes’ training can make a difference, according to a report in the photography website PetaPixel.
Dr Katie Gray, the study’s lead researcher from the University of Reading, wrote that the increasing realism of AI-generated faces presents real-world risks.
“Computer-generated faces pose genuine security risks. They have been used to create fake social media profiles, bypass identity verification systems and create false documents,” Gray says in a University of Reading press release. “The faces produced by the latest generation of artificial intelligence software are extremely realistic. People often judge AI-generated faces as more realistic than actual human faces.”
“Our training procedure is brief and easy to implement. The results suggest that combining this training with the natural abilities of super-recognizers could help tackle real-world problems, such as verifying identities online.”
