Snapchat is exploring how facial recognition technology could be used to protect the privacy of its users.
A new patent from the smartphone messaging company, issued by the United States Patent and Trademark Office on July 19, is entitled: “Apparatus and method for automated privacy protection in distributed images.”
It discusses how sophisticated face-recognising technology could be used to scan new photos, and then — depending on who it detects — it could either refuse to allow the image to be shared without permission, or even automatically edit the image to protect the subject’s privacy.
It’s a novel use for facial recognition tech, which is a frequent subject of concern for activists due to its potential to invade people’s privacy and facilitate surveillance. Snapchat’s patent paints an alternative possibility: The tech could actually be used to enhance people’s privacy.
A Snapchat spokesperson declined to comment.
You can read the full patent below.
Using emojis to defend people’s privacy?
United States Patent number 9,396,354 was first filed with the USPTO (United States Patent and Trademark Office) on May 27, 2015. It has two inventors listed: Evan Spiegel, cofounder and CEO of Snapchat; and Robert Murphy (better known as Bobby), cofounder and CTO.
The language used in the patent filing is dense and technical, as patents often are — but it details an interesting concept. For example (emphasis ours):
“A method executed by a computer, comprising: receiving an image from a client device; executing a facial recognition technique against an individual face within the image to obtain a recognised face; applying privacy rules to the image, wherein the privacy rules are associated with privacy settings for a user associated with the recognised face; and distributing a privacy protected version of the image, wherein the privacy protected version of the image has an altered image feature, wherein the privacy protected version of the image is distributed to an ephemeral gallery … [and so on]”
In plain English?
- Snapchat could build a facial recognition database of its users, and these users would be able to save certain privacy settings about how they want to let photos of them be shared.
- Then, when a photo of a user is taken, the feature would scan it to see which Snapchat users it can detect.
- If a detected user has certain privacy settings enabled, then the feature will limit how the photo can be shared. It might only be able to be sent to other users who are in the subject’s “social network.” Or the subject might get sent a message with the image in, asking for their permission so it can be shared.
- Or, if the photo-taker does not have permission to share a photo of someone, the feature will automatically modify the image before it is sent — blurring out the subject’s face, or replacing it with an emoji, or an icon, or “a brand.” (Another option is to block the sharing of privacy-protected images altogether.)
It protects user privacy — but only in certain ways.
At this point, it’s important to note that just because Snapchat has filed a patent doesn’t mean that the feature will necessarily make it into the official app. Big tech companies file thousands of patents a year, many of which never see the light of day.
But the patent filing is interesting for a number of reasons.
For a start, it suggests that Snapchat is actively exploring facial recognition technology.
The app is already notable for its filters (called “lenses”) that map users’ faces and overlay images and animations on top of them — but these use “object recognition” rather than facial recognition tech.
“[It’s] an algorithm designed to understand the general nature of things that appear in an image. It lets us know that a nose is a nose or an eye is an eye,” Snapchat says on its Privacy Centre website. “But object recognition isn’t the same as facial recognition. While Lenses can recognise faces in general, they can’t recognise a specific face.”
The patent shows that while Snapchat is currently only making use of object recognition tech, it is at the very least thinking about the possibilities of facial recognition tech too — and may well be actively developing it.
It’s also an unconventional approach to protecting privacy, and one that arguably has its downsides.
Yes, it protects you from unauthorised photography through the app. But, in the process, it gives Snapchat a detailed facial recognition database on its users.
Snapchat offers pretty poor targeting options to advertisers compared to the likes of Facebook or Google. The nature of a search engine or social network means you can target people incredibly precisely — while Snapchat only gives limited options like age, location, and gender.
The Los Angeles startup doesn’t seem to crave this kind of fine-grained detail, however: CEO Evan Spiegel has explained plenty of times that Snapchat’s aim is to avoid the kind of “creepy” ads that internet users often hate.
That said, if you can recognise exactly who is any given photo, you can theoretically target ads far more precisely.
The patent discusses the possibility of monetising the feature: “Facial recognition operation may [also] be used to add content, such as a sticker, text or an advertisement, which do not necessarily protect privacy or hide identity, but do provide an enhanced image,” it reads.
However, it’s not explained how these advertisements would work. We don’t know whether they would simply augmented a user’s face, like existing filters — or if they would actively target ads towards the specific users in photos. The latter would be unexpected, given Spiegel’s aversion to “creepy” ads.
We do know that Snapchat has been thinking about using object recognition tech to deliver targeted ads, coupons, and filters. A recently-published patent details how the tech could be used to detect where a subject is, or what objects are in a photo, and serve up ads accordingly.