Opt-outs, removal, and the limits of both

People who find their face somewhere it should not be usually want it gone. The word gone hides three different outcomes: not findable by face, not findable by Google, and not on the site at all. Each needs a different request to a different party. The opt-out routes directory and the NCII removal directory hold the current forms; this entry explains the system.

1. Step 1

   Face search opt-out

   Ask each engine to suppress matches of your face. Scope: that engine's results only.

   ›
2. Step 2

   Search engine de-listing

   Ask Google or Bing to drop a page or image from results. Scope: that search engine's results only.

   ›
3. Step 3

   Host removal

   Ask the website, or its host, to delete the photo. Scope: the actual copy.

   ›
4. Step 4

   Repeat

   Copies elsewhere need their own requests; hash-blocking services reduce re-uploads on participating platforms.

What does a face search opt-out do?

An opt-out adds your face to a block list. The engine computes an embedding from the photo you submit and suppresses any result that matches it. The photos stay on their source pages and in the engine's index; they are hidden from output. Consequences follow: the engine now holds your embedding on purpose, the block works only as well as the match threshold, and a photo of you that scores below the threshold still appears. Routes differ: PimEyes and Reversely.ai require an anonymised ID scan, FaceCheck.ID verifies with a selfie, Lenso.ai takes a face photo, FaceSearch takes a request by email with a face photo. Each engine must be asked separately.

What does a Google removal do?

Google offers several forms. Results about you removes pages that show personal contact details. The request for non-consensual explicit imagery removes sexual images of a person from results. The outdated content tool refreshes results for pages that have changed. In the EU and UK, a right-to-be-forgotten request under GDPR Article 17 can de-list results for a person's name. All of these affect Google's index; the page itself remains online and reachable by direct link, and other search engines are unaffected.

What does removal from the host do?

Only the host can delete the copy. The legal levers are: a DMCA takedown if you own the copyright (you usually do if you took the photo, and usually do not if someone else did), a platform's own impersonation or NCII policy, a data-protection erasure request where the GDPR or a similar law applies to the host, and statutory duties such as the US TAKE IT DOWN Act (48-hour removal of non-consensual intimate images by covered platforms from May 2026) or the UK's Online Safety Act. Hosts outside these regimes, or anonymous sites, may ignore requests; the next step is the hosting provider or domain registrar, found through WHOIS.

How do hash-based blocking services fit in?

StopNCII.org (adults) and Take It Down (under 18) compute a perceptual hash of an intimate image on the victim's own device and share only the hash with participating platforms, which then detect and review uploads that match. This prevents re-uploads of the same file on those platforms. It does not find existing copies, does not cover platforms that have not joined, and does not match a new image of the same person.

Where do all three stop?

A working order of operations

1. 01 Find every copy you can: face search plus reverse image search, saving URLs and dates.
2. 02 Hit the host first for the copies that matter most, choosing the strongest lever available (NCII duty, then copyright, then policy).
3. 03 De-list from Google and Bing so the remaining pages stop surfacing on your name.
4. 04 Opt out of each face engine so future searches on your face return less.
5. 05 Register hashes with StopNCII or Take It Down if the images are intimate.
6. 06 Recheck quarterly; the web does not stay fixed.