How old is this person? Let the AI guess from a photo
Upload a photo of anyone and our AI tells you how old they look in a few seconds. If the photo shows a group, it reads the biggest face.
Your photo is deleted once it is read. Your result is deleted within 20 hours. Never used to train.
The short answer
You can get a fair guess of how old someone looks from one clear photo. Our AI finds the face, reads the signs that change with age, and gives one number: the age the person appears to be. It is the same model behind our how old do I look test. It guesses the look, not the birth date, so treat it as a second opinion, never as proof.
How to guess someone's age from a photo
It takes three steps and works on a phone.
- Get the photo. Save the picture, or take a screenshot of the post you saw. Crop it so the face you care about is the biggest one in the frame.
- Upload it above. The AI finds the largest face and reads it. If it finds no face, try a sharper or closer crop.
- Read the number. You get the age the person looks. If you know their real age, type it on the result screen to see how many years younger or older they look.
One tip matters more than the rest: use the clearest photo you can find. A blurry screenshot of a screenshot loses the fine detail the model reads, and the guess gets worse.
"How old is she?" Settling the comment section
Posts that ask "how old does she look?" or "guess his age" are everywhere. Someone shares a photo, a few hundred people guess, and the answers spread over twenty years. Then the real age comes out and half the comments were wrong.
An AI guess is a neutral referee for that game. Write your own number down first, then upload the photo. The model has no idea who the person is, what they do for a living or what the caption claims. It only sees the face in the picture. That makes it a good check on the story the caption is telling you.
Why people get someone's age so wrong
Most of us are worse at guessing ages than we think, and the misses follow a pattern. Researchers who study age guessing have found that people judge smiling faces as younger, and that they tend to guess too young for older faces (Voelkle and colleagues, 2012). People also guess better for faces close to their own age.
A photo adds its own tricks:
- Light. Soft light from the front hides lines. Hard light from above deepens every shadow and adds years.
- Filters and editing. Beauty filters smooth the skin. Many viral "she is 60 and looks 30" photos are edited, and the face no longer shows its real texture.
- Makeup, hair and beards. They change the parts of the face we read for age, in both directions.
- Angle and distance. A wide phone lens close to the face changes its shape. A photo from below or from the side hides the jaw and the eyes.
The AI is thrown by many of the same things. The difference is that it has no mood and no manners. A friend might be kind and shave off a few years. The model just gives the number.
How the AI reads a face
First it finds the face in the photo and crops to it. Then a deep learning model looks at skin texture, fine lines, the area under the eyes and the overall shape of the face. It compares those cues with patterns it learned from a large set of faces of known age, and returns its best guess.
On a good photo the guess is usually within a few years of the real age. On a hard photo, with a filter, a big smile or poor light, it can miss by more. We explain the model and its limits on the how it works page.
Use it fairly
Guess the age of people who are fine with it: friends, family, yourself, or public figures in published photos. Do not use it to judge strangers, and never use it to decide whether someone is old enough for something. A face guess is not proof of age, and it can be off by years for any one person.
Which photo gives the best guess
The original, not a repost
Every repost and screenshot loses detail. Use the sharpest copy you have.
Face toward the camera
A straight, relaxed face lets the model see the eyes, mouth and jaw.
Crop to one face
In a group photo the biggest face is read. Crop to the person you mean.
Unfiltered if you can
A smoothing filter pulls the guess down. Expect a younger number from edited photos.