How to Spot AI-Generated Photos and Deepfakes on Dating Apps
For years, the standard advice for catching a fake dating profile was simple: reverse image search the photos. If they belonged to someone else, you would find them. That advice still matters — but in 2026 it is no longer enough on its own. AI image generators can now produce a unique, photorealistic face that belongs to no real person and matches nothing online, and real-time deepfakes can fake a live video call. The good news: AI fakes still leave tells, and a few deliberate checks will catch the overwhelming majority of them. Here is exactly what to look for.
Why AI Photos Changed the Game
Traditional catfishing relied on stolen photos — pulled from a stranger's Instagram or a model's portfolio. That method has a weakness: the images exist elsewhere online, so a reverse image search can expose them. AI-generated faces don't have that weakness. Each one is synthetic and unique, so searching for it returns nothing — which means a “no matches found” result is no longer the reassurance it used to be.
Worse, a scammer can now generate dozens of distinct, believable personas in an afternoon, each with its own consistent-looking face. The volume of fake profiles and their believability have both gone up at once. So the goal shifts from “is this photo stolen?” to a broader question: “is there a real, specific person behind this profile at all?”
How to Spot an AI-Generated Photo
AI image models have gotten remarkably good at faces — but they still stumble on the details around the face. When a profile feels a little too polished, zoom in and look closely.
Hands, Ears, Teeth, and Jewelry
These are where generators still fail most often. Look for extra or fused fingers, mismatched or “melting” earrings, teeth that blur together, and glasses with asymmetric or warped frames. The fine edges of accessories are a frequent giveaway.
Warped Backgrounds
AI faces are convincing; the world behind them often isn't. Check for garbled text on signs, doors and railings that bend impossibly, blurry “melted” objects, and background people with missing or distorted features.
Hair, Edges, and Skin
Strands of hair that dissolve into the background or merge with clothing, an airbrushed-looking seam where the face meets the hairline, and skin that is flawlessly smooth in every single shot are all soft signals — especially in combination.
Too Perfect, No Candids
A whole gallery of studio-perfect, head-on, evenly lit shots — with no casual snapshots, no friends, and no specific recognizable locations — is a pattern. Real people have messy, contextual photos from actual life.
The Consistency Test
A single AI image can be flawless. A set of images of the “same person” is much harder for a generator to keep consistent — and that is your best lever. Across all of their photos, check whether it is genuinely the same face:
- Do freckles, moles, or scars appear in some photos and vanish in others?
- Does their apparent age drift between shots?
- Is a tattoo present in one photo and gone in the next?
- Do eye color or the shape of features subtly change?
Stolen photos are internally consistent, because they are of one real person. AI-generated sets often aren't. Inconsistency across a gallery is one of the strongest signals you will get that the face isn't real.
Deepfakes on Video Calls
A live video call has long been the gold standard for confirming someone is real — and it is still one of the best tools you have. But real-time deepfakes now exist, so the call itself deserves a little stress-testing. Current face-swap tools still struggle with occlusion, extreme angles, and unpredictability. On a call, casually ask for things that break them:
Ask for Movement the Fake Can't Follow
Ask them to turn their head fully to the side (deepfakes often “slip” at a full profile), slowly wave a hand in front of their face (passing something over the face is where the illusion tends to tear), or press a finger to their cheek or nose (the swap frequently warps where the hand meets the face).
Watch the Edges and the Audio
Look for shimmering or flickering face edges, hair or glasses that glitch on movement, lighting on the face that doesn't match the room, an oddly static background, and any lag between their lips and their words.
You can frame all of this casually — “show me your room,” “wait, you've got something on your cheek” — without it feeling like an interrogation. A real person does these things instantly. A deepfake puppet hesitates, dodges, or visibly breaks.
How Intently Helps
Intently offers verified profile badges that confirm a member passed identity verification with a photo and ID — a meaningful signal that a real, specific person is behind the profile. A badge isn't a substitute for the checks above, but it raises the floor. For the full pre-meeting routine, see our guide on verifying someone's identity before meeting in person.
What to Do If You Suspect AI
You don't need certainty to protect yourself — you only need to slow down. If something feels off:
- Don't accuse — ask. Request a casual live action: a quick video call, or a spontaneous photo of them doing something specific right now. A genuine person obliges easily.
- Reverse image search anyway. It still catches stolen photos, and a no-match combined with other red flags is itself informative.
- Watch the timeline. AI personas are often paired with classic romance-scam behavior — fast affection, reasons they can't meet, a push to move off the app. Our guide on spotting and avoiding romance scams covers the full pattern.
- Report the profile. Reporting helps Intently investigate and protects the next person who matches with them.
Quick AI-Photo and Deepfake Check
- Zoom in on hands, ears, teeth, and jewelry for artifacts
- Check backgrounds for warped text and melted objects
- Look for candid, real-life, location-specific photos
- Compare all photos for consistency (freckles, age, tattoos)
- Reverse image search every profile, every time
- Have a live video call and ask for a spontaneous action
- Cross-check for a social presence with real history
- Slow down and report if anything doesn't add up
The Bottom Line
AI hasn't made safe online dating impossible — it has just moved the goalposts. The instinct that protects you is the same as ever: verify before you trust, and treat reluctance to do something simple and live as the red flag it is. Spend two minutes zooming in on the photos and a few minutes on a real video call, and you will catch almost every AI fake long before it can become a problem. The technology behind the fakes is new; the discipline that beats it is not.
Date With Confidence
Intently's verification badges and platform-level safety tools help you focus on real connection — not on wondering whether the person you're talking to is real.
Join Intently