Google Glass Getting A Face Recognition App This Month, But It Won’t Get Google’s Blessing

Google Glass use cases are many, but one that inevitably comes to mind is facial recognition. Google already does a lot with reverse image searching and identifying faces in photos, so it would not be such a leap to imagine it doing something like comparing the faces of those you meet at networking events to publicly available photos from Google+ and other sources to make sure you never again forget a name. But Google has forbidden that kind of software in the official Glassware app store. Still, startup Lambda Labs and its founder Stephen Balaban are building that software anyway, for installation via sideloading.

That workaround means the app, called FaceRec, will only ever make it onto a fraction of Glass devices, and a Google spokesperson had this to say when contacted for comment as confirmation it’ll never get broad distribution through any official channels:

As our Glass Developer Policies make clear, we will not be approving any Facial Recognition Glassware.

A subset of the Explorer crop can’t add up to many installs, but that’s exactly who it’s intended for, Lambda tells Forbes. The app works by storing a record of every face that a user encounters while wearing glass, on a cycle that refreshes to capture new faces every ten seconds. In this early version, it can’t ID faces in real-time, and doesn’t have a reference database from which to draw. Instead, like with iPhoto and other services, you can tag pictures with names so that they’ll be recognized the next time you see them. Users can also roll their own script for mining data from their Facebook network for automatic identification, but it’s not built into the product since it violates Facebook’s rules of usage.

The first version of Lambda’s Glass facial recognition app might be limited, but it’s a first step to something more on par with what we might expect from sci-fi examples, where you glance at someone and get a profile of them, shared interests and more provided via a heads-up display. Which is great, because getting to know people the old fashioned way through conversation and a gradual deepening of mutual understanding is for the birds.

Seriously though, there does seem to be a general level of anxiety around the idea of Google Glass and facial recognition. But over time we’ve proven ourselves to be quite changeable on the definition of what is and isn’t acceptable when it comes to how much information we share with others via the web, and facial recognition could become something that people grow more comfortable with time. It definitely has a range of positive possible use cases, including for treatment of genuine medical conditions like prosopagnosia or the aftermath of strokes.

Google may eventually relax its privacy restrictions to make this kind of app officially supported on its Glass platform, but Lambda is also building its own Android-based wearable device called the “Lambda Hat” that will be available for pre-orders Friday. This and other platforms developed outside of Google likely won’t carry similar strictures about face recognition tech, so Balaban’s concept of a world where we can know people just by looking using computer vision might come to pass regardless of Google’s reservations, and the serious privacy implications such a concept entails.

This may be a particularly interesting example of unauthorized Glass software, but software outside the bounds of platform restrictions is nothing new. Apple has a far-reaching and active iOS jailbreak community, after all, and Android devs have created many apps that can be sideloaded but don’t make it into the Play Store. Glass is bound to play host to a few of those as well, but novel technology makes for novel takes on what constitutes ‘out-of-bounds’ software. None of these unauthorized apps really make it beyond outlier or curiosity status, unless policies change and they gain access to official channels, but they can still be worth watching as barometers of what users find interesting and/or acceptable in specific examples of mobile software.