In a study attempting to observe virality in real time, two Microsoft engineers, Corom Thompson and Santosh Balasubramanian, used Microsoft's recently released Face detection API to create How-Old.net and track its usage in real time.
Suffice it to say the study went well, and what came out of it was a fun service that lets you upload an image in order to determine the age and gender of the people in it. Let's show you how to put the HowOldRobot to the test and see whether it's accurate or not with your personal photos.
On the How-Old website, you have the option to test it out with provided images, but you don't know how old those people really are, so select "Use your own photo." For the best results, find a high-resolution picture where each face is clearly visible and facing forward.
Though the gender detection isn't working too well, the API was able to closely determine the age—give or take a few years—of the men in the photo. The female's face is partially covered and wearing glasses so it rendered some inaccurate results in that aspect.
Not everyone may be too eager to jump on their desktop to test this out, so navigate to How-Old.net on your phone and use the mobile-friendly version. You can even take a photo with your phone , or select one from your device or cloud storage services.
Let me warn you now: some of the results may offend you. I've been tagged as being over forty a few times (I'm 26), and had a male friend labeled as a female (though I won't hold that against Microsoft's technology, he does have very gentle features).