The New Face Race

by Erik Sass, Sep 28, 2012, 5:56 PM

• Comment
• Recommend (1)
• Tweet
• 

Redpepper’s Facedeals are the latest in facial-recognition innovations. Will this incarnation be the one to blast through consumer fears?

Well, Minority Report arrived a lot faster than most of us probably expected. Who knew it would have a Southern flavor?

Lost in the Olympic hubbub this summer was some interesting news from a small creative advertising and media agency in Nashville called Redpepper, which is testing an innovative new marketing platform in Nashville combining social media, digital out-of-home, brick-and-mortar retail venues, and facial recognition. 

In the Redpepper system, called Facedeals, in-store cameras scan the faces of customers who have opted in, then use facial-recognition software to match them to their Facebook profiles, check them in to that location, and automatically deliver targeted promotions to their smartphones based on their interests and preferences, as expressed in their “Like” history. In other words, the system knows who you are (or at least, what you told Facebook about yourself) and uses that information to try to sell you stuff.

Perhaps inevitably, the announcement of Redpepper’s experimental system provoked expressions of skepticism (to put it mildly) in the trade and consumer press. The word many people reach for is “creepy,” and it’s not hard to see why: Technology that can “see” you, identify you, and connect you with other forms of personal information summons up images of the two-way video screens used by “Big Brother” in 1984 and, yes, the ubiquitous surveillance in Minority Report.

There are some obvious differences, of course: For the system to work, the consumer must opt in to the Facedeals app on their Facebook account and connect their Facebook account to their smartphone; the scope of the video monitoring is limited to certain venues; and the man behind the curtain is just trying to sell you something, not finger you for the prison camps, making all those totalitarian analogies a bit overdrawn. Still, few people will be surprised if Facedeals and similar systems have to contend with reluctance stemming from our collective fear that, basically, robots are spying on us.

And Facedeals does indeed represent a considerable advance (if that’s the word) over other systems that use facial recognition for marketing in public places. In other iterations — for example digital signage created by Quividi or TruMedia, or digital out-of-home networks created by companies like Provision and InWindow Outdoor using Intel’s Audience Metrics Suite — none of the information collected is used to actually identify individuals. Although some of these systems might be able to match names to faces if provided with a database of known images, they instead tend to be limited to more general information, for example gleaning demographic data like age, gender and race, as well as gaze-tracking to determine where passers-by are looking, in order to target advertising more efficiently. On the other hand, none of these systems require the user to opt in, and even without explicitly identifying someone, they can still create anonymous profiles which allow them to track behaviors over time.

The real reason Facedeals, currently limited to a handful of shops in Nashville, has aroused so much interest is what it represents in terms of future capabilities. Redpepper has done the right thing by requiring consumers to opt in — but there’s nothing to stop some other agency or DOOH network from adopting a more aggressive model that takes more liberties connecting identities to other sources of personal information, including social media or consumer databases.

For example, a considerable proportion of Facebook users still leave some personal information publicly visible on their profiles: A study published in February of this year by researchers at the Polytechnic Institute of New York University found that just 33 percent of Facebook users opted to hide all personal information on their public profiles (including age, high school name and graduation year, network, relationship, gender, interests, hometown and current city). Any one of these pieces of information, together with a publicly available name and a single photo, would be enough to allow facial recognition software to match individuals with personal characteristics for purposes of delivering ad targeting — whether they’ve opted in or not.

And that’s not all: Given the capabilities of online analytics engines and behavioral targeting platforms, there’s nothing to stop marketers from further connecting an individual in public to their online browsing history. It doesn’t take much imagination to see that combining an individual’s online behavioral data and their precise coordinates in the real world would yield a bonanza of actionable insights to marketers: For example, someone posting about their home-renovation project on Facebook and researching power tools online could receive targeted ads — via mobile, digital signage, or both — when they’re near a hardware store. It’s one thing to see disarmingly relevant ads delivered to your Web browser while you’re online; how will consumers react when the ads start appearing on digital screens at the mall or out on the street?

Lest anyone doubt that this is technically feasible, just consider that almost a decade ago, in 2003, some facial-recognition software could compare an individual face to a database of known faces at the rate of 70 million images per minute; at that rate, a system using facial-recognition technology could theoretically scan an individual’s face and compare it to every Facebook user in the U.S. in a little over two minutes — and the software’s capabilities have likely advanced in tandem with computing power over the last decade. Additionally, some software can now recognize moods and even health conditions, which could impart an extra level of creepiness to ads targeted using that information, anonymously or otherwise.

In short, there’s plenty of reason to be leery of where technology is taking us, both as a profession and society more broadly. Playing devil’s advocate, however, it’s not impossible to imagine ways to make such systems work without alienating the public.

The most important element, of course, remains the opt-in: Hopefully, most marketers will recognize that, even if it it’s technically feasible, any attempt to match social media profiles to individuals in public spaces without their explicit permission will likely end in disaster. In any event, increased consumer awareness of basic Web privacy practices may, over time, limit the amount of publicly retrievable information about individuals floating around online in the first place.

Presuming that opt-in systems become more common, there remains the question of incentives: How can marketers persuade individuals that they should allow themselves to be tracked in this potentially unsettling way? One obvious way is discounts and special offers, which could be extended to cover a wider range of venues and product categories. How many people would agree to opt in for facial recognition in return for, say, an across-the-board discount of 5 percent off everything they buy — from cars to consumer electronics to packaged goods to gas to airline tickets, and so on? Conversely, once someone has opted in for facial recognition, a single advertiser could reach that person across multiple platforms: For example, a series of video ads forming a unified narrative could be delivered to an opted-in individual starting on their PC or mobile device, then migrating to DOOH screens in public spaces.

But all this remains hypothetical, at least for the time being: It’s important to remember that the rules governing even the first wave of facial-recognition technology — meaning simple, non-identifying uses like those described above — are still very much up in the air. In a letter to the Federal Trade Commission published in December 2011, the Center for Digital Technology, which has pioneered the policy debate about facial recognition, pointed out that “federal laws — and nearly all state laws — do not provide American consumers with basic privacy protections when it comes to biometric information collected for commercial purposes online or offline.”

For its part, the industry hasn’t been entirely idle when it comes to self-regulation. In the same letter the CDT commended two industry organizations, the Digital Signage Federation (DSF) and the Point-of-Purchase Advertising Institute (POPAI), for drawing up voluntary guidelines intended to govern the use of facial recognition by DOOH networks and place-based advertisers. These include requiring consumer opt-in for directly identifiable information; forbidding the collection of biometric information from minors; and providing notice to consumers when biometric data is being gathered, even if it is anonymous. But the CDT also notes that “there is no overarching set of privacy standards covering all or even most commercial uses of facial recognition, and the overall compliance rate with existing privacy standards related to facial recognition is unknown.”

Just as important from a marketing perspective, given the potential risks to brand reputation and image, it’s also doubtful whether the general public is currently even aware of the capabilities offered by basic facial-recognition technology. As this awareness grows and public opinion about facial recognition gradually coheres, it will doubtless drive the legislative and regulatory process forward — but given the current low level of awareness, the eventual outcome is all but unforeseeable.

• Comment
• Recommend (1)
• Tweet
• 

Be the first to comment on "The New Face Race"

Leave a Comment