That’s what droves of Internet users are saying after a new website takes a stab at guessing their age. On How-Old.net, Microsoft’s facial recognition software analyzes a photo of you and tries to determine your age. Web users everywhere seem to love it – until the app places them in the AARP crowd long before they’ve celebrated the big 4-0.
Tweets abound that blast the bot for adding years onto people’s actual age. (It added 5 years to actress Jennifer Lawrence’s true age of 24.) But sometimes, it makes us feel young again. There are plenty of self-congratulatory Facebook posts by people judged to be from the “can I see your ID?” crowd, when in fact they’ve been absent from it for a long time. (It guessed age 29 for me – a 43-year-old dad who doesn’t even use cold cream).
The site’s lackluster track record serves as proof that machines have a long way to go to match humans in intelligence.
Besides offering fresh fodder for your newsfeed, what’s the benefit for Microsoft to invest in facial recognition technology – especially when preliminary reviews wind up in critics’ crosshairs?
It’s an auspicious element of artificial intelligence. Deep learning involves artificial neural networks, providing data from audio, images, and more. Every time a user taps into the app – regardless of accurate results – they provide data. The more data it acquires, the more accurately it can predict outcomes. Microsoft isn’t the first to employ deep learning – but it is the first to provide it in products for developers to use.
The big picture
How-old.net incorporates elements of Microsoft’s Project Oxford, an initiative that fuels the company’s natural data understanding. The group also delves into speech recognition and image categorization. Imagine an app that melds all three functions.
Your Face and the Marketplace
Your face can be more than a digital playground for age estimation. Facial recognition technology can scan you in stores while you shop to customize your experience – and potentially help a vendor to access your personal data.
Microsoft plans to install versions of its Kinect camera with facial recognition software in stores to track three metrics:
WHAT’D YOU GET? | What gets picked up? Vendors can track which products wind up in your cart.
WHERE’D YOU GO? | Where do you head after you choose an item? This data can show retailers opportunities for additional sales.
WHO ARE YOU? | The software can create a shopper profile. It evaluates a shopper to craft a face template – a numerical code derived from facial measurements. The software links clothing size and shopping history, and is accurate even if you wear sunglasses or a hat.
Retailers recognize a gold mine of data up for grabs. They could match images taken on the sales floor with your social media profile. Leading providers of this technology don’t disclose their clients, averting possible backlash similar to what Nordstrom experienced when it tracked customers by smartphones.
The plus side: Intelligence allows retailers to customize offers. For example, casinos can have a high roller’s favorite cocktail to them minutes after they walk in.
The minus: The blurred boundary of privacy created when data is mined without consent. Who’s accountable if that data falls into the wrong hands, or gets misused?
The United States is one of the few developed nations without a data collection/sharing law in place. The American Civil Liberties Union decries its use as inaccurate and a violation of privacy.
Face Recognition and Law Enforcement
Marketing advantages are just part of facial-recognition technology. Security services use it, and notify retailers if known shoplifters, past employees or felons enter the premises. (British police even apprehended a suspect by matching his face to images of close relatives.)
Opponents of data harvesting don’t have a leg to stand on with no official policy or regulations in place. The FBI has launched the Next Generation Identification program, which will amass 52 million photos by the end of 2015.
The Electronic Frontier Foundation, a non-profit organization focused on civil liberties in the digital world, claims the FBI will obtain 4.3 million images for noncriminal activity, such as employer background checks.
Facial recognition use for security purposes, such as automated account login or gated-community admittance, is a nice modern convenience. What if hackers breach this data the way they did credit cards recently?
Today’s trendy age-guessing app doesn’t answer that question. It just sheds a little light on the path facial recognition technology could take.