1. The lighting doesn't match the scene
Real photos have light that comes from somewhere a window, the sun, a lamp. Most AI generators apply lighting that looks generically "nice" instead of physically accurate. Shadows fall in the wrong direction, or there are no shadows at all. Your eye notices this before your brain can explain why.
2. Skin looks airbrushed, not photographed
There's a specific kind of smoothness AI defaults to too even, too clean, missing the small imperfections that make skin look like skin. Real photos have texture: pores, slight unevenness, the way light catches a cheekbone. Remove all of that and you get a face that looks rendered, not photographed.
3. Backgrounds are vague or impossible
Look closely at the background of most AI images and something is usually wrong text that isn't real text, architecture that doesn't make sense, objects that blend into each other. Your brain registers the inconsistency even if you can't immediately name it.
4. The styling doesn't match a real moment
AI-generated outfits and poses often look like a mood board, not a moment. A real photo captures a person existing in a specific second mid-laugh, mid-step, hair slightly out of place. AI content tends to look posed for no reason, frozen in a moment that never happened.
5. There's no editorial point of view
This is the one most people miss. A good photo, AI or not, has a style decision behind it a color grade, a mood, a reference point. Generic AI tools skip this entirely and just generate "a nice image," which is exactly why it looks generic.
What to look for instead
The fix isn't more detail or higher resolution. It's curation. Tools that apply an actual aesthetic a specific lighting style, a specific mood, a specific editorial reference produce results that read as intentional, not generated. That's the difference between content that gets scrolled past and content that gets saved.