Microsoft showcases face detection API with a web site that guesses your age based on a photo, which quickly goes viral
it's a lot of fun to play around with Alejandro Alba / NY Daily News : Microsoft launches How-Old.net to guess a person's age, Internet reacts Greg Kumparak / TechCrunch : How Old Do You Look? Microsoft Built A Robot That Tries To Guess Your Age Adnan Farooqui / Ubergizmo : Microsoft Launches A New Website To Guess Your Age Brittany Hillen / SlashGear : Microsoft's ‘How Old’ age estimator reaps Internet mockery Tweets: Steve Kovach / @stevekovach : The Microsoft age detector is the new bunny with sign
Context & Ripple Effects
In 2015, Microsoft turned its face detection API into a party trick: How-Old.net guessed users' ages from a photo, went instantly viral, and drew as much mockery for wrong answers as praise for the tech. It was a template — a lightweight consumer demo doing the marketing work of an enterprise SDK.
A decade later the arc has inverted. The same age-and-emotion inference moved into products like the Sprinkles captioning app, then out again — Microsoft's Responsible AI Standard led it to remove Azure's facial analysis tools entirely in 2022. Meanwhile photo inference has grown more powerful and more unsettling elsewhere, with ChatGPT users using o3 to pinpoint where photos were taken.
First-order effects
- Microsoft gets massive free publicity for its face detection API, but the viral spread is driven partly by mockery of inaccurate guesses, tying the brand to unreliability.
Second-order effects
- The demo normalizes uploading personal photos to corporate servers for trivial entertainment, seeding the privacy expectations that later force reversals like the Azure facial-analysis removal.
Third-order effects
- If the pattern holds, playful perception demos precede regulatory and reputational blowback: capabilities that launch as toys get withdrawn or restricted once deployed at scale, pushing vendors toward explicit Responsible AI governance.
The trend: Face and photo analysis has traveled from viral novelty to governed liability, with Microsoft's own trajectory — demo, productization, withdrawal — marking the industry's maturing stance on biometric inference.