Have you ever scrolled through Zillow and thought, “Wow, what an awesome agent!” then questioned if all of those wonderful reviews are real? Plot twist: at times, they’re not. Sometimes, actually, they’re typed by a bot who likely doesn’t even own property.
What’s Going On With Zillow Reviews
A recent report from Originality.ai analyzed more than 10,000 reviews from Zillow agents and determined 23.7% were likely AI-generated. That’s almost 1 out of every 4 reviews. Yep, one out of four. So when someone gushes about their “stress-free home buying experience,” there’s a pretty good chance it was from a robot who has never even attempted to get pre-approved.
In 2019, the percentage AI-generated was just 3.63%. That’s a 558% jump in just a few years. If that was a stock, you’d call it a rocket ship. But since it’s fake reviews, let’s just call it a headache. Zillow, for its part, hasn’t confirmed any plan to label reviews as AI or human. Basically, you’re still left playing detective while trying to figure out if “Best Realtor EVER!!!” was penned by Karen in Ohio or a sleepy algorithm in the cloud.
Why This Matters (Beyond Zillow)
This isn’t Zillow’s alone clumsy issue. AI is getting really good, and scary good, at writing in human style. A recent research paper showed that humans can’t reliably tell the difference between fake and real reviews anymore. On average, people detect authenticity only about 50.8 % of the time, basically by chance.So, we’re roughly as bad at identifying AI reviews as we are at identifying terrible dating profiles.
This is to say that the next time you hear “best customer experience ever” from a restaurant, a hotel, or an agent, you may indeed be reading the robot version of it.
How to Tell If a Zillow Review Is Probably Fake
So how do you find the bots in the wild? Here are some hints, because alas, Zillow does not provide lie detectors:
- Too repetitive wording: When several reviews read like they were cut and pasted by a homeschooled teenager doing their homework under the gun, that’s an indicator.
- Vague compliments: “Excellent experience, very professional, always replied.” Nice, but did they buy your house or pet-sit your cat?
- Overly euphoric writing: Actual people gripe about something. Too good? Too artificial.
- Timing patterns: Ten five-star reviews suddenly pop up on one Tuesday? Sure, John.
- Generic profiles: No picture, no background, just “JD” gushing about life-altering service. Completely credible.
Even AI detection gets this one wrong occasionally, which is kind of funny. Picture the bot that generates the review also passing the test to show that it’s human. Similar to witnessing a magician clapping at their own trick.
What Platforms and Legislation Are Doing
Thankfully, this circus isn’t going completely unchecked. The FTC is rolling out new rules banning fake reviews, including AI ones. What does this mean? If you’re caught faking, you’re paying.
Meanwhile, Zillow has launched something called a “Fair Housing Classifier” to prevent AI tools from discriminating in chats. Nice start, but how about a “Fair Fake Classifier” that calls out reviews written by ChatGPT’s cousin?
Other sites are already testing verified reviewer labels and proof of purchase prior to review publication. Consider it an online review bouncer. No receipt, no comment. So the next time you scroll through Zillow and read all those 5-star reviews singing an agent’s “unparalleled professionalism,” keep this in mind: one of those reviews was probably written by a coffee-sipping bot. The reviews aren’t leaving, but now you know how to read them with one eyebrow raised. And if you can’t differentiate? Flip a coin; you’ll be just as accurate as an expert.
