Is this image AI generated?
An AI detector can provide a clue, but checking a picture means looking at its origin and context too. Work through these steps before deciding what the image shows.
Updated October 28, 2026 · AI Checker
1. Find the original source
Start with the post or page where you found the image. Look for a credited photographer, an original upload, a publication date and a caption. Follow attribution back to the creator when possible. A repost with no source gives you less evidence, not proof of AI generation.
Keep the original file if it is available. Record the source URL and the context you want to verify before making edits or taking screenshots.
2. Search for earlier appearances
Use a reverse-image search to find other versions and pages using the picture. Google Lens can return similar images and websites containing them; see Google’s image search instructions.
Compare captions, dates and crops. An older appearance can show that a picture has been reused in a new context. No search matches does not establish that it is new, authentic or AI-generated.
3. Examine details without treating them as proof
Look closely at text, repeated patterns, reflections and the relationship between objects. Use inconsistencies as questions to investigate. Blur, perspective, editing and compression can also make real images look unusual; realistic details do not rule out generation.
4. Review metadata and provenance
If available, inspect the file’s creation and editing information. Content Credentials use the C2PA provenance standard to describe origin and editing history. Provenance is a different kind of evidence from a detector’s visual assessment.
Missing information is inconclusive. AI Checker reports provider-supplied metadata evidence and does not independently verify credential signatures. Do not treat a displayed metadata field as an authentication certificate.
5. Use a detector as another signal
Try the AI image detector with the original JPG, PNG or WebP. Read the report’s mode, signal and limitations. A likely-AI label calls for closer review; a low signal does not prove a camera original. If results disagree, keep that uncertainty in your conclusion.
A worked example of careful interpretation
Suppose a reposted photograph looks unusual, has no metadata and receives a high AI signal. That combination is a reason to investigate, not a verified origin. Search for the earliest publication, ask for the original file and compare its caption with other evidence. Until the source is established, “origin unverified” is more accurate than claiming the image is definitely fake. This is an illustrative scenario, not a measured detector test.
What can you conclude?
Separate three questions: whether a file was generated, whether it was edited, and whether its caption is true. One tool cannot answer all three. For consequential decisions, corroborate the source and keep an explanation of what remains unknown.
Read how AI detector accuracy works, or check an image. Before uploading sensitive material, review our privacy policy.