AI Photo Detector for Profile, Dating and Listing Photos

Use this AI photo detector when the question is about someone you might meet, pay or trust: a dating match, a new contact who messaged you first, a seller on a marketplace, a host or landlord with a listing. Add the photo above and you get an answer with a probability range and the reasons for it.

It runs the same checks as our AI image detector on the home page. Your browser reads the photo's Content Credentials and camera metadata first, without uploading it. If those don't settle the question, the photo goes to a detection model that judges the pixels.

The oldest profile-photo trick needs no AI at all: someone steals real photos of a real person. A detector will correctly call those Likely a real photo, and the profile is still fake. So treat the answer as a clue about how the image was made, not about who you are talking to.

To check identity, run a reverse image search to see whether the same face turns up under another name, and ask for something that can't be prepared in advance, such as a short live video call. Do that before you send money or share anything personal, whatever the detector says.

Calling a real person's photo AI can cost them a match, a sale or a rental, and they have no easy way to prove you wrong. That is why an AI photo detector should never claim 100%: ours shows a range, and says Can't tell when the signals are weak.

In our test, neither Sightengine (our main engine) nor Hive (our backup) flagged any of the 50 real photos, across every type in the table above. That is encouraging, but 50 photos is a small sample and can't rule out a false-alarm rate of a few percent. Two other engines we tried on small samples did flag real photos: AI or Not flagged 1 of 10, and WasItAI 2 of 5.

An original file from a phone camera normally records the make and model in its EXIF metadata, and that is one of the first things an AI photo detector reads. In our set, 37 of the 50 real photos named a camera, from iPhones, Pixels and Galaxy phones, and none of the 50 AI images did. Some phones go further. Google says Pixel 10 attaches Content Credentials to every JPEG it captures, and in our set, real photos from a Pixel 9 and a Pixel 10 carried camera credentials.

The catch: most social platforms strip metadata when a photo is uploaded, so a picture saved from a profile will usually have none. Missing metadata is normal and is not a sign of AI. Metadata that is present can be edited, so it supports a conclusion rather than settling it.

Each answer describes how the image was made, never who the person is or what they intend. Read it alongside what you already know about them, and act on it gently, because a wrong guess lands on a real human being.

• Likely a real photo: the pixels look camera-made. It does not confirm the person's identity or that the photo is recent.
• Likely AI-generated: a reason for caution, not proof. Ask for a fresh photo or a quick video call before you meet or pay, and don't post the result or accuse the person publicly.
• Edited with AI: the file itself says an AI tool changed it. Google Photos, for example, marks pictures touched by Magic Editor or Magic Eraser this way. Lots of real people use these tools, so it means the photo was retouched, not that the person doesn't exist.
• Can't tell: no usable signal either way. Rely on the other checks above.

With a rental, a used car or a second-hand phone, the risk is a listing whose pictures don't show the real thing, and an AI photo detector only sees part of that risk. Generated room shots are one version of that; photos lifted from someone else's old listing are another, and a detector won't catch those.

Ask for a short video that pans around the item or the room, or a photo with a note bearing your name next to it. Our set included 6 real product photos and neither main engine flagged any of them, but we didn't test listing photos of homes or cars specifically.

Only part of the story. It can tell you whether a photo looks generated. Many fake profiles use stolen real photos, which a detector will rightly call real. Pair it with a reverse image search and a short video call.

No. We don't contact anyone, and the check never touches the app where you found the photo. We keep only a hash of the image and the verdict, plus a small thumbnail only if you share the result and leave the thumbnail option on.

Check the original file from your phone, not a screenshot or a copy saved from an app: the original keeps your camera details and, on recent Pixels, its Content Credentials. We didn't test heavily filtered selfies, so we can't say how often filters set off a false alarm. If someone acted on a wrong label, show them the original file.

No. Photos saved from social apps have often had their metadata removed on upload. Treat missing metadata as no information and let the other signals decide.