Upload a portrait to Google Lens and the results show pages that carry that exact picture, shops selling the clothes in it, and a grid of visually similar images. Lens is not built to return a different photo of the same person, and Google has said in writing why. It can trace where a picture came from, while the person in it stays unnamed unless they are a notable public figure.

What Google says Lens does with a face

Google answers this most directly in the FAQ for Lens inside YouTube Shorts. Asked whether viewers can use Lens to search for people read in November 2026, says: "We don't use any biometric facial recognition with Lens. However it's possible that we may show results for notable public figures."

The reasoning behind that sits on Google's "Our approach to facial recognition" page. It separates face detection, which only establishes that a face is in an image, from face recognition, which works out whose face it is. It also separates authentication, confirming that a person is who they claim to be, from mass identification, picking someone out of a database without their consent. Google writes that this is "why we've been so cautious about deploying face recognition in our products, or as services for others to use."

So a photo of a well-known actor can come back with a name, because Google allows results for notable public figures, while a photo of a private person comes back with pictures and no name.

Google does group faces in one product. Google Photos has a Face Groups setting that sorts the photos in your own library by the faces in them. Google Photos Help says only you can find the private face labels you choose, even when you share those photos, and turning Face Groups off deletes the groups and the face models built to make them. The grouping stays inside your own library.

What Lens matches instead

Google's Search Help page for iPhone lists what a search with an image can return: AI Overviews, search results for objects in the image, similar images, and websites with the image or a similar image. Put a face photo through it and each result type answers a different question.

• Websites with the image. If the photo was copied from somewhere, this is where the original page and its reposts appear. If you are asking "who is this?", this list is the useful part.
• Objects in the image. The glasses, the shirt, the guitar, the building behind the person. Google built Lens for this job, from products to plants, animals, artwork, murals and landmarks.
• Similar images. Photos that resemble yours as pictures. With no face recognition involved, resemblance means pose, framing, lighting and setting, so the grid can fill with strangers who were photographed the same way.

One practical trick follows. Google's help page notes that selecting a smaller area of the image gives more specific results. Drag the box around the jacket to shop for it, or around the building to identify the street. Keep the whole frame when you want the photo's source.

Tracing where a photo came from

If your real question is where a picture came from, Lens is a good first tool. This routine keeps the results useful:

1. Start from the best copy you have. A photo saved from a chat app, or a screenshot of a screenshot, has lost detail at every step.
2. Search the full image first, with nothing cropped.
3. Scroll past the shopping and similar-image results to the pages that contain the image itself.
4. Open each of those pages and note who posted the photo, under what name, and when.
5. Save the URL and a screenshot of any page that uses the photo under a different name.

A photo that appears under several unrelated names is a reason for caution about the account where you found it.

When a face search is the better tool

A face search locates the face in your photo, converts it into a numeric description of facial geometry, and compares that against faces on publicly reachable pages. Background, clothing and colour grading are discarded, which is how it can connect a studio headshot with a holiday snapshot of the same person. FaceSearch has a plain guide to how a face search differs from a reverse image search, with the two compared side by side.

Use it when your question is about a person: whether the profile photo of someone already messaging you appears elsewhere under another name, or whether your own photos are being used on accounts you do not control.

A face search reads public web pages only. It cannot see inside private accounts, closed groups, messages or apps that sit behind a login. A match is a lead to open and check at its source page, never proof of identity, and an empty result proves nothing either way about whether a person is real. It is also for adults only, so never run the face of anyone under 18.

Lens and a face search, side by side

Before you search a face

1. Decide whether your question is about the photo or the person. The photo goes to Lens, the person to a face search.
2. Search only an adult's photo, and only for a reason you could explain, such as an account that has contacted you or a picture of yourself.
3. If you use Lens often, open your Google Account, go to Data & privacy, then Web & App Activity, and check whether Visual Search History is ticked.
4. Treat every match as a lead and open it at the source before drawing a conclusion.

If either search turns up your own photo on an account you do not run, take a screenshot that shows the URL and the date, then report the account through that platform's impersonation form.