How to Spot AI-Generated Images on Social Media
Check the source before the pixels. On social media the fastest tells are not visual at all: who posted first, whether any credible account carries it, and what the caption wants you to believe. After that, check provenance and run the visual checks. If you cannot verify an image that would spread fast, wait instead of sharing.
Why it is harder here
Platforms re-encode images on upload. That strips C2PA Content Credentials, which are the clearest evidence of origin. Invisible watermarks like Google SynthID are built to survive that, but only the vendor and some platforms can read them, so a normal viewer cannot. Screenshots strip even more. By the time an image reaches your feed, much of the evidence is already gone.
Quick checks for feed images
- Source first. Who posted it, when, and does the account have a history? A viral image from a week-old account is a warning.
- Reverse search. Google Lens or TinEye shows whether the image circulated earlier, elsewhere, or with a different caption.
- Zoom in. Check hands, text, reflections, repeated furniture, and warped lines. Our seven checks cover the details.
- Compare the caption. A real photo used with a false story is as misleading as a fake one.
Use the platform labels and provenance
Some platforms now add an “AI-generated” label, and a few read C2PA credentials. Treat a label as helpful but incomplete, because unmarked tools and screenshots slip past all of them. When you can get the original file, inspect it at contentcredentials.org/verify. OpenAI now attaches both Content Credentials and Google SynthID to its generated images, and a public verification tool at openai.com/verify checks those marks, though only for OpenAI-generated content.
Score it, then decide whether to share
If you have the image file, run it through the free AI image checker for a 0-100% score. Use the score alongside the source and the reverse search, not instead of them. A high score plus a thin source is a reason not to share. The same reasoning applies when you are the one posting, so the guide to Is This Image AI? is worth keeping handy.
Common questions
Why do AI images look obvious on my feed but fool people elsewhere?
Context. A striking image that fits a story spreads before anyone checks it. The visuals may be fine, but the claim around the image is doing the work. Check the source and the earliest post before you judge the picture alone.
Do platforms remove AI metadata?
Many re-encode or strip metadata on upload, which removes C2PA Content Credentials. Pixel watermarks like Google SynthID are designed to survive that, but only the vendor and some platforms can read them. A screenshot removes even more.
How do I check an image I saw on social media?
Screenshot it only as a last resort. Better, open the original post and save the image file, run a reverse image search, and score it with a detector. Then check who posted it first and why.
Should I share a shocking image before verifying it?
No. Sharing is amplification. A quick check takes a minute: reverse image search, look for provenance, read the source. If you cannot verify it, say so or wait.
Check before you share
AI Detector scores any image in seconds and never stores it.
