Signs to check with your own eyes
Start with the parts generators still handle badly. Look at hands and fingers, teeth, ears and jewelry, where you often find an extra digit, a melted edge or an asymmetry. Read any text in the image, on signs, labels or clothing, since garbled letters are a strong tell.
Then step back and check the physics. Do the shadows all fall the same way? Does a reflection match what should be in front of the mirror or the window? Skin that looks too even, backgrounds that blur into nonsense, and repeated textures that tile are all worth a second look.
When to bring in a detector
Your eyes catch obvious fakes, but a clean generation can pass a human check. This is where an AI media detector helps: it scans the file for statistical traces that generators tend to leave, such as diffusion-model patterns or pixel-level artifacts, and returns a probability with a confidence level.
Danaya offers a free detector for exactly this. You can drop in a photo, video or audio file and get a result, with one free check a day and no account needed. Treat it as a second opinion that sees things the eye cannot, not as the final word.
A likelihood score is not proof
A detector reads a finished file, so it is always a step behind the newest generators. As models improve, their traces shrink, a clean fake can slip through, and a genuine photo that was heavily compressed or edited can be flagged by mistake. A score tells you how suspicious a file looks, not what actually happened.
When you need certainty rather than a probability, the honest answer is to prove authenticity at the source. Danaya can seal a photo the moment it is captured, binding a SHA-256 fingerprint of the file, the GPS location and a server timestamp. Anyone can then verify that seal from a public link. Use detection to screen photos you receive, and sealing to prove photos you produce.