I am looking at a situation where the pictures that are posted on a listing are maybe not telling the whole truth. Now, we have a new situation in our whole market today, across the entire industry, of it being extremely easy now to generate as opposed to take photos.
When we list a property, we want to list it in a way that’s very marketable. And so, we go out and we dress up the home, you know, we stage it, we try and do good lighting, we try to make it look the best we can, and then we take photos to sort of portray the best possible face of this home.
We’ve got great pictures of the inside with the right furniture. We’ve got great pictures of the outside, maybe at the golden hour, great view. You take them at the right time in the spring when the buds are blooming, and all of that is an accurate representation of the home at the moment that you take that photo.
Now, what we’re having trouble with now is that there are a lot of tools that allow you to modify those images to make it look better. We’ve always had that with Photoshop, you could adjust the colors and you could even remove undesirable elements in the edge of the photo.
When Does Editing Become Dishonest?
You could modify the image in certain ways, editing the photo. But the degree to which you do it changes whether it’s honest or dishonest. There’s this idea we have called puffing, such as when you say, “This is the best view you’ll ever see.” That’s something that’s not really provable. It’s not necessarily lying or dishonest because it’s an opinion.
But when you’re looking at photos of a property now, this isn’t an opinion, right? When most people look at a photo, they believe that what they’re seeing is going to be an accurate representation of the property. Now, they’ll probably recognize that it’s the best possible light or it’s a great setup, but they believe the images are not going to tell lies, right?
But with these new image modifying tools, these AI tools that can completely redo images, now we’re starting to see images that are not telling the truth, and they’re being posted in a listing of a property. So, for example, if we look at staging the home, if you bring in furniture that wasn’t there before, it still represents an honest picture of what somebody might set up the home to look like. We know that the furniture is not going to stay when we buy a home, and so if the furniture leaves, that’s fine. But when we do this AI creation with these photos, sometimes it changes real physical elements of the home.
It might remove a staircase from the perspective on a living room where there was one in the background, or if you’re taking an outside shot, it might change some of the outside characteristics. It could change what the roof looks like, it might make it look like it’s made of something else. It could change which trees are planted in the yard. Those are big permanent structures, and the AI just adds them or removes them really easily. If the agent isn’t watching, they might let something slip through that tells a dishonest story about the property.
How Can AI Tools Be Dishonesty?
Earlier I mentioned taking photos in the spring when things are in bloom. It’s beautiful, and that’s something you would want to do. But let’s say you listed the home in the winter. So, you use a little bit of image modification, maybe an AI tool, and now you’re showing what it’s going to look like in the springtime, which people want to see.
But because you don’t have a real picture of the property, you’re having it generated. That basically means it’s being invented, created out of an idea, not out of the real property that’s there when you have the AI create a springtime photo.
Even the word “photo” implies something that’s real as opposed to a generated image. We’re going to have to change a lot of the dialog and the words we use to match these new things that are happening out there. But if you generate a springtime photo of that home, it likely will insert flowers that don’t exist on the property, bushes and blooms that might not be there.
It might show the trees that aren’t on the property or remove trees that would have been on the property. These generative images are very smart, but they’re not necessarily telling the truth. They’re telling what something could look like if it were a different property, not if it’s the same property.
Who Should Be Held Responsible?
So as agents start to use these tools, we need to be careful that we are conveying an honest image. If I take a home and I list it, I take some photos and I create some photos with the AI, but then I post those photos on the listing, who is responsible to ensure that those are accurate?
Who is responsible? Who is culpable? Who is the one that’s going to be left holding the bag? The reality is that I am the one to put up the listing. It’s got my name on it. That means that I’m endorsing the content on there as if it’s accurate. I am the professionally licensed real estate agent or broker that is saying that this is accurate, and so I can’t pass the buck to the AI tool I was using because I could use any tools.
I could use Photoshop and invent a picture. I could have one of my staff go and take pictures of the wrong house, and I could post pictures of the wrong house. Whose fault is that? They might have contributed to it, but I’m the one who’s responsible for my listing. So, if I post images that do not accurately reflect the property, I am the one who is responsible.
And if I’m a real estate sales agent, it may be that my real estate broker is also responsible for that content accuracy. So I could get my broker in a lot of hot water by posting images that have been modified by tools such as Photoshop or AI, tools that present an inaccurate picture of the property. It can become a level of fraud if we’re purporting that there are trees on a property when they’re not, or that there’s not a staircase when there was.
Ensuring The Accuracy of Listing Images
There’s a level of modification of the photos that can be allowed because it still represents what the home potentially might be, and there’s a level where it’s no longer accurate. That’s starting to crop up more and more, because these image modification tools are so easy to use that we forget that we need to take that extra part of our responsibility as a practitioner, as a licensee, where we signed to say we would be honest and ethical and that we would portray an accurate advertising image of our properties to the public.
It’s up to us to make sure that those tools are used accurately. Sometimes even just taking pictures that are too old of the property and then trying to use those today where the foliage has changed or where the paint is peeling now or the roof has been redone can be deemed dishonest if they don’t reflect the current status of the property.
I think a lot of realtors could even take that advice to look at the photos they’re using of themselves in their marketing. Asking themselves, “Is that still what I look like today?” That’s kind of a common issue, isn’t it?
Using AI Tools In Image Descriptions
There’s another element of this. What if you’re not modifying the image itself? What if you’re just identifying what’s in the images? Like, there’s a tool in the Multiple Listing Service that I use regularly. It looks at all my photos and will write a description of each room based on what it automatically identifies. And again, that is a generative tool which is very useful, but it’s up to me to make sure that those descriptions are accurate.
The AI might write it up and say this is a bedroom, when it’s really just a great room or something, but there was something in there that made it think there was a bed. The tool might do that, but I am still responsible to make sure that text is accurate.
You know, it’s funny. With an image sometimes we think, “Oh, well, that bush could be there, right? They could plant that bush.” But if you wrote in the text, “Has a gardenia bush,” and then it doesn’t have it, that’s clearly fraudulent. In the photo it should be the same, but for some reason we don’t have that same standard.
Psychologically, we look at those things differently, but they are the same. Both things are representing what the property is, and both things should be honest and accurate. And so, if the tool is coming up with inaccurate room descriptions, that’s up to me to fix those. It’s up to me as the real estate agent who is posting these listings to make sure that they’re accurate and that it’s an honest advertisement for the property.
Now, some of the terms we use in real estate have sort of changed over the years. Like recently we started saying things like “owner suite” instead of old terms like “master bedroom,” or things like that. I think the most preferred term I’ve heard for that is “primary bedroom” now, in an attempt to go beyond past stereotypes or avoid terms that might be exclusive or discriminate against anybody.
And so, we want to make sure we use those new terms that are open and clear to be inclusive of everybody who participates and who wants to buy the home.
Humorous Modifications to Listing Images
As a final thing, I would say that there have been some usages of AI in real estate images that have been intentionally humorous.
For example, I saw a listing where every room in the house had a sneakily inserted Godzilla in the picture. Now, if people looked at that, it might be so obvious that we say, “Okay, that’s humorous. That’s not fraud, that’s not dishonesty, that’s an obvious attempt at humor,” but it is also technically something that’s inaccurate, right? That there was that creature in the home.
So using humor is fun sometimes, but we need to make sure that in the advertising we do of our homes, when we’re using these smart tools which can be very useful, we still need to remember that the responsibility falls on us, the practitioner, the licensed person who is taking responsibility for the accuracy of the advertisements. We need to make sure that we are presenting a clear picture to the public, for our clients, and that we’re treating everybody with respect and honesty.
So keep an eye on what you post under your own name, and let’s make it the best it can be.