Answer first: You can no longer trust a face or a smooth conversation on its own, because AI can now generate both. To get more confidence that someone is real, stop judging by photos and words alone and ask for a live, spontaneous video call. Someone who is real and interested will usually say yes. Someone who keeps making excuses is telling you something. Old-school catfishing meant stealing photos from a real person's social media. It left clues, like reverse image search hits or a story that did not add up. AI-era deception is different. The face may belong to no one. The messages may be written by a chat model that never gets tired, never slips up on grammar, and always says the right thing. This post is about that new layer, and how to stay grounded when fakes look better than ever. Why this is harder now For years, the best advice was to reverse image search a profile photo. If the same picture showed up on other accounts or stock sites, you had your answer. That advice still helps, but it is weaker now. AI can create a brand new face that has never appeared anywhere online, so a reverse search comes back clean even when the person is fake. Chat has changed too. In the past, scam messages were often full of odd phrasing or copy-pasted lines. Today an AI tool can hold a warm, flirty conversation in fluent language, at any hour, with endless patience. The polish that once signaled a real, thoughtful person can now be produced by software. That does not mean everyone is fake. It means the old shortcuts are no longer enough on their own. Signs a photo might be AI-made AI images have gotten very good, but they still slip up, especially in the details. Look closely instead of just swiping past. No single sign proves a photo is fake, but a few together are worth pausing on. Skin and lighting that look too flawless, almost airbrushed, in every single shot. Odd details at the edges: extra or blurry fingers, teeth that blend together, jewelry or glasses that warp. Backgrounds that melt, repeat, or contain text that is garbled and unreadable. Hair or clothing that fades into the background instead of having a clear edge. A set of photos that all feel the same, with no candid shots, no friends, and no everyday moments. Real people have messy, ordinary pictures. A profile that looks like a polished ad shoot, with no normal life around it, deserves a second look. Signs the chat might not be a real person AI-written messages can feel charming yet strangely hollow. The tone is warm, but the substance is thin. Watch for chat that stays generic and dodges specifics. Real conversations have texture. A real person remembers what you said three messages ago, reacts to your jokes, and shares small, specific details about their day, their town, or their work. AI-driven or scripted chat often stays vague. Ask a concrete question, like what part of the city they live in or what they did last weekend, and notice if the answer is oddly smooth but empty, or if it dodges the question and pivots back to compliments. Another tell is pacing that feels off. Replies that arrive instantly at all hours, or that pour on affection very fast, can point to a script or a bot rather than a busy human being. Fast, intense attention early on is meant to lower your guard. Behavior signs that matter more than looks Some of the strongest warning signs have nothing to do with pixels. They are about what the person does. These patterns show up in both AI-driven scams and older ones, and they still deserve your full attention. Excuses or flat refusals when you suggest a live video call. There is always a broken camera, a bad connection, or a reason it can wait. Pressure to move off the app quickly, onto text or another messaging app, where there are fewer safeguards. Fast escalation of emotion, calling you special, soulmate, or their whole world within days. Any request for money, gift cards, crypto, or help with an investment. This is the clearest stop sign of all, no matter how real they seem. If money enters the picture, the conversation is over. No genuine new match needs your bank details. Not now, not later. One of the strongest checks: a live video call If you take one thing from this post, take this. The most practical way to ground yourself in reality is a live, spontaneous video call before you meet in person. Deepfake video exists, but running a convincing live fake in real time, while reacting naturally to your questions, is much harder than posting a fake photo or sending AI-written text. Keep it simple and unscripted. Have a normal chat. Ask them to wave, turn their head, or hold up a hand. A real, comfortable person can often do this naturally. Faked or manipulated video often lags, glitches at the edges of the face, or the person keeps finding reasons to avoid the camera altogether. Steady, repeated refusal to ever appear live is one of the loudest signals you can get in the AI era. What verification can and cannot do Profile verification is a real help, and it is also not magic. When an app checks that a person is a real, present human before they can fully take part, it raises the cost and effort of running a fake. That extra friction can discourage low-effort scam and bot accounts. It makes deception harder. What verification cannot do is read someone's heart. A verified person can still be dishonest about their intentions, their relationship status, or their feelings. Being real is not the same as being trustworthy. So treat verification as a strong foundation, not a finish line. Keep paying attention to how someone treats you over time. How HoopFrog is built for this We built HoopFrog around this exact problem. Our approach is verification-first, so being a real, present person is part of joining, not an afterthought. We support live video calls before you meet, so you can see and hear a real human before investing more of yourself. There is no swiping, which keeps the focus on genuine connection instead of endless snap judgments on photos. And when something looks wrong, our team can review escalated reports. None of this is a guarantee about anyone's character, and we will never pretend otherwise. It is a set of safeguards that raise the cost of fakery and give you more real signal to work with. Ready to date with more confidence in the AI age? See how HoopFrog puts real, verified people first, and lets you meet on video before you meet in person. Real. Verified. You.