CBSN

How-Old.net will guess your age, test your self-esteem

Upload if you dare.

A new site from engineers over at Microsoft called How-Old.net will take a photo of you and use facial recognition software and machine learning to guess your age. Prepare to be amazed. Or devastated. Or possibly flattered.

CBS News' Gigi Stone Woods and Vladimir Duthiers, along with CNET senior editor Jeff Bakalar got a little taste of all those responses when they put the "How Old Robot" to the test with their own photos. The site guessed Bakalar's age within four years, just outside the one-to-three year range of accuracy experienced by many of the tens of thousands of people around the world who have tried the software. Duthiers was unpleased with his estimate; Woods took the Fifth.

Microsoft is keeping the details of exactly how the software works under wraps, but it leverages the power of existing facial-recognition technology and machine learning that uses the data it amasses to improve performance.

"The more people use it, the more people upload photos," Bakalar explained, "it gets closer and closer" to correctly guessing age every time.

Microsoft is certainly not the only major company looking into new ways to put facial recognition to use. "Facebook is all over that," said Bakalar. "They definitely target ads based on the people in the photos that you're looking at."

On the back of its DeepFace facial recognition software and massive archive of photos, the social network is getting astoundingly good at recognizing people in pictures and automatically tagging them. An article published in Science in February cound that DeepFace is now "as accurate as a human being at a few constrained facial recognition tasks."

  • Amanda Schupak

    Amanda Schupak is the science and technology editor at CBSNews.com