Detect AI-Manipulated Evidence in Refund and Damage Claims

Upload the photo to the checker on this page. You get a likely verdict, AI-generated or human-made, with a score out of 100. If it comes back as likely AI, do not refuse the claim on that alone: ask for a new photo or a short video taken now, and compare the item with what you shipped.

IsItAI.com detects whether a photo was made by AI, and that includes photos edited in AI chat apps like ChatGPT or Gemini, which redraw the whole picture when you ask them to add a crack or a stain. Tools that change only a selected area keep the rest of the original photo, so a human-made result is not proof that nothing was changed.

Use the result to decide which claims a person should look at, not as the decision itself. Genuine damage photos can look unusual, and no detector is right every time. A likely AI result is a good reason to ask for more evidence.

Videos can be faked too. The AI video detector scores one frame every 2 seconds, up to 59 seconds per check, and shows where AI appears.

Yes. The API takes images by upload or URL, up to 10 MB each, and checks up to 20 images in one request, so a support or returns tool can check every claim photo as it arrives.

Not on its own. Photos sent through chat apps, email or a screenshot usually lose their metadata, so missing camera details prove nothing. ChatGPT and some phones add a signed Content Credentials label to the images they make or edit, and we show that label when the file still has it. Without it, the AI score still checks the photo itself.

Yes. Results are private by default. A public result page is created only when you choose to share it. See our privacy policy.