If your Facebook feed is anything like mine, it’s flooded with wave after wave of AI-generated aesthetic house content: picturesque tiny homes, sleek art deco bathrooms, ornate bohemian interiors, cathedral-like modern kitchens, and more.
At first glance, it’s hard to tell if these photos are real, and it’s getting to the point that even a close look may be inconclusive. Very rarely, one lone commenter or two will leave a note that it’s AI in the comments, but that’s just one drop in a sea of other remarks like: “Beautiful!” “Wow! I wish I could live there.” “Where is this?”
This has driven me up a wall. But why? What’s the harm? This isn’t political misinformation. We’re just talking about pretty pictures of homes. Why do I feel such a strong need to let others know that it’s not real? Why do I feel so at sea when I can’t quite tell if a picture is AI or not?
Philosophical Concept of Porn
I think the answer might be found at least partly in the philosophy of porn—porn in the sense of food porn, house porn, inspiration porn, and even moral outrage porn. As C. Thi Nguyen and Bekka Williams define it, home porn in this generic sense would be images of rooms, houses, or other spaces created to provide immediate pleasure without you having to actually use the space or take care of it. As they explain: “Porn is free to go to extremes because its consumers don’t have to deal with the complications of the real thing. With porn, we get to skip the hard part.”
You can enjoy the image of the elaborate AI-generated sink without having to think about cleaning all the crevices. You can enjoy the image of the vast bohemian roomscape without having to think about cleaning all the fabrics or maintaining the impossibly placed house plants. In these images, vibes almost always trump practical design. They are unobtainable fantasies.
Epistemic Issues (epistemic has to do with knowledge)
I think the main reason why the AI house porn bothers me is that it’s passing off fantasy as reality. There is no clear indication on these pages that they are AI-generated, and this very much feels like a new kind of scam that I need to learn how to detect.
Because these pictures are AI-generated, they can be produced exceptionally fast and their quality is improving rapidly. So, it’s hard to keep up with and develop methods of detecting AI at an individual level, let alone a societal level. And if you really want to fact-check everything, that takes a lot of effort. You can’t keep yourself clean in a sea of bullshit.
Even as a certified AI sleuth (I spend far too much time trying to determine if the homes Facebook shows me are real or not), I can’t always tell. This has shaken my epistemic confidence more generally: When Princess Catherine announced she had cancer, I had genuine doubt at first as to the veracity of the video. What if it was an AI generated deepfake?
When I point out to friends that the AI aesthetic content they repost isn’t real, I don’t mean to be the party pooper. Yet, they tend to be disappointed when they discover it’s AI generated. I think this is telling. We do want these places to be real. We do want to visit them. We do want to experience living in these beautiful rooms. There is a betrayal when we want to bring the fantasy to reality.
The commenters on the original posts also generally treat them as if they are real—some note design flaws as if these were practical spaces and others sincerely ask where these images were taken. Some commenters want to visit the home or find out more information, and they can’t. (Or, a scam website is linked in the comments as if it provided more information.)
Interestingly enough, it’s the fact that so much of this is porn winds up to be the key to detecting it—it’s going to extremes. The houses are too pretty to be true.
If we get swept away by the vibes, porn that disguises itself as reality can warp our expectations and give us cartoonish ideas about the world. When we can’t tell the difference between reality and AI porn, we set ourselves up to be conned and exploited by impossibly good fantasies that can be generated in seconds.
In some ways these problems aren’t new. This is just another wave of Facebook feeding deceptive images and posts into people’s feeds. We’ve seen this happen before with fake news and moral outrage porn (the focus of Nguyen’s and Williams’ article), which uses oversimplified stories and narratives to cleanly separate the good guys (you) from the bad guys (them) and allow you to feel morally superior.
What’s new is that AI speeds up the rate of bullshit production and the rate at which shady content can be re-tailored to appear more legitimate. If we start getting more AI generated political misinformation, then we’re going to be in for a real ride in the 2024 presidential election.
This is not just a Facebook problem. Google recently had to change their algorithm to combat AI generated spam websites that were ranking well in the algorithm—I came across one or two while searching for health information. They were near the top of google results, and I could only tell it was AI once I started reading the prose.
Aesthetic Issues
On the aesthetic side of things, AI porn just isn’t as good as the real thing. It’s missing certain historical, personal, design, and narrative connections and values.
I think what’s so frustrating to me is that the AI generated porn feels like one step beyond an HGTV-like design porn that makes renovations out to be easy and that standardizes design taste into a set range of commodifiable styles. The AI images present an escapist fantasy in a world where design doesn’t really matter and there are no practical limitations. That can be all fine and good when we know what it is. But the deceptive nature of these Facebook pages makes it much harder for us to know what’s fantasy or reality.
This feels especially dystopian when home ownership is out of reach for so many right now and when others are becoming homeless due to higher rents and a lack of housing availability. The world we actually live in has real design problems that need to be solved, and good city design and house design is often sacrificed in the name of capitalist greed. It’s all too easy to forget those problems and indulge in aesthetic AI porn.
But if we lose our sense of the elegance of designs made for real people who live in these real spaces, we can easily despair at the unavailability of these extreme aesthetics or become fooled by the appearance of good design that’s rotten underneath (I’m looking at you flipped houses with grey or greige luxury vinyl planks!).
My favorite design content to watch is made by people who are renovating homes themselves and revealing all of the hardships, tedious steps, and ups and downs in restoring beauty to older buildings. Take The 2nd Empire Strikes Back and DK Dreamhouse. Both creators deeply love the design and histories of the homes they are working to live in, and there is a tangible sense of progress with each step forward. I fantasize about fixing up an old house one day, and these creators help to show me how that could be a reality. That is a beautiful gift.
There can still be a place for pretty AI content that people seek out for the vibes. But, if there’s any kind of truth claim involved or some financial incentive (This is a real place you can rent! This is a real pie and recipe you can make! This is a real outfit you can buy!), we enter the realm of possible scams and deception. And that’s solidly where these mass-posted images currently lie.
Tinfoil Hat Time
I don’t know yet what these ubiquitous AI-generated pages are trying to achieve. Are they just learning what gets engagement? What looks real? What people like? Is this a way to improve engagement with a slowly dying Facebook?
At least two different pages had the same link to what appeared to be a scam website once I clicked on it. This makes me think that they’re probably run by the same entity. I also would not be surprised if the initial engagement on these posts is boosted by bots themselves. This all feels shady.
If your reaction to this article has been to say “stop clicking on and interacting with these posts because it’s messing up your algorithm!”, I’ve tried. The instant I started rejecting the AI house content, Facebook started feeding me AI-generated images of pies and cakes with AI-generated recipes. And, my friends have been sharing the same house content I’ve been seeing on their feeds as well—this is a ubiquitous problem, and I’m not sure how to personally escape from it.
Bonus: A Quick, Incomplete Procedure for Determining if It’s AI
Look at the photo.
Zoom into repetitive patterns. Are the windows consistent? Is the tile pattern regular?
Look for too many objects. Are there too many lights on the front of the house? Is there a massive number of roses climbing on the tiny home?
Think about the practicality of the space. Would anyone actually design a kitchen with exceptionally difficult to clean windows?
Watch out for hyper-realism. Are the colors too clean and crisp? Is it just impossibly beautiful?
Look at the other information provided by the post.
Check that the poster is reputable and reliable. Is this a new page you’ve never seen? Do they provide any specific information about who they are on their page?
See if any information that can be fact-checked has been provided. Is there historical information about this place? Is there a link to a reputable website?
If you can’t tell from the photo or page, reverse Google image search.
See if there are other results from reputable websites. If you can find more information about history or location, it’s likely real. If there’s only one photo on a shady website, it’s likely fake.
The more AI content you identify, the better you’ll likely get at spotting it at the first step. But some real places look fake sometimes, so don’t be overconfident at a glance.
Have you encountered this kind of content? What's your reaction to it? Let me know!