Protection Guide · 5 min read · July 21, 2026

How to Check Possible Sources of AI-Generated Intimate Images

Learn how to check a possible source of an AI-generated intimate image and what face search and AI detection can each tell you.


How to Check Possible Sources of AI-Generated Intimate Images

If you suspect an intimate image uses your likeness, separate two questions. Where does the image appear? What evidence is there that it was generated or altered? Each question calls for a different check.

Start with that page. Record the URL, publication date if visible, account name, and any information the post gives about the image. If you have a known original, compare the two images. Look for differences that can be tied to a source, such as a changed background or a reused crop.

When you only have a photo of yourself

Use a clear, ordinary face photo you own or are authorized to use. PrivacyLeak can look for possible face matches in supported adult-related sources. Full source links and details require an unlock. Check any relevant result in context and keep it private while you verify it.

For example, suppose a search returns a similar face from an adult-related domain. That result gives you a possible page to investigate. The person could be a lookalike, and the page may contain a different image than the one you heard about. A search that returns no results leaves open the possibility of content outside the supported sources.

Check how the image was made

An image detector assesses an individual image for signs of AI generation. Face search compares facial similarity across indexed sources. To assess an image, look for a reliable original and use any detector output as one piece of evidence. Compression and editing can also change the details a detector sees.

How to compare a suspected image with a known original

Use a private copy of the original you have permission to view. Check whether the face angle, hair, expression, and lighting line up. If only the face looks similar, there is little evidence that the two images are related. If multiple details line up in the same places but the scene differs, the images may be related. Check the dates and source pages before concluding which one came first or how the change was made.

For example, a person may recognize an expression from a public profile photo in a suspicious image. Matching the expression alone is weak evidence. An earlier source containing the same crop, hairstyle, and lighting gives you a stronger comparison. Keep both URLs and record which details you could verify.

What different findings mean

Finding Next step
The same image appears on two pages Keep both URLs and compare publication dates and attribution.
A similar face appears in a different image Inspect the context and leave the identity unconfirmed if details are unclear.
An AI detector flags the file Compare with the original and assess how the copy was compressed or edited. Keep the detector result as a separate clue.
Search returns no relevant page If you have a suspected URL, document it directly. Try another clear authorized face photo if you still need to check supported sources.

If you confirm a concerning source

Keep the URL and relevant context, then report the page through the platform’s intimate imagery or impersonation process. For guidance on the next steps, see how to respond to a deepfake intimate image.

FAQ

An ordinary authorized face photo can be used as the input. If you already have a source URL, save it separately. It gives you a direct place to check even if the search does not return that page.

Ask for the page URL or the account and platform where they saw it. Avoid forwarding the image to verify the claim. You can look for possible appearances using an ordinary authorized photo and treat anything you find as a lead until you can verify the source.