Back in September of last year, I introduced you to SceneTap, a service that uses webcams to provide a dynamically updated count of how many people are currently in a bar, restaurant, etc…along with estimates of male-to-female and age ratios, thereby implementing not only human face detection but also rudimentary face recognition (gender and age, but not specific-individual identification). The information is, perhaps obviously, of interest to folks interested in meeting others, therefore in assessing the "situation" at various establishments ahead of potential visits. It's also potentially valuable to establishment management, in assessing the success of "ladies night" and other similar promotions. And of course, SceneTap aspires to eventually also serve up ads, along with the statistics, to cellphones.
At the time I first mentioned the company, SceneTap was predominantly based in Chicago, but last week it expanded to San Francisco, in the process bearing the brunt of even more controversy than it had previously experienced. The backlash was predominantly fueled by a writeup in SF Weekly, a local newspaper, even though the coverage explicitly noted:
Bars place special facial detection cameras inside venues, which pick up on facial characteristics to determine approximate age and gender of the bar crowd. All your personal information remains anonymous, and nothing about you or your face is stored long-term.
After assessing the predominantly negative feedback left by SF Weekly readers (likely fueled by the publication's labeling of the service as "creepy"), several establishments that had been planning on launching SceneTap services changed their minds at the last minute. SceneTap initially published an "open letter to the city of San Francisco" which, among other things, strove to reassure concerned residents with refreshingly candid statements such as:
Here’s the thing – there are no videos or images stored at any time. Once the data is triggered, the images are overwritten, deleted, gone. There are no tapes. There is no video feed either. No one can go to www.scenetap.com and see what is happening. It’s all data and numbers – that’s it. And since we’re only focused on the door, you’re free to do keg stands and dance like Bernie or hit on that bartender all you want – we do not track you in the venue.
Not everyone was appeased. In a statement to Ars Technica, for example, Rebecca Jeschke, a digital rights analyst for the EFF, commented "It is in fact creepy! Looking at the privacy policy, they say they don't keep video or stills, but are silent on if they keep the measurements and other data they collect in order to make their conclusions about gender and age. That's a big question for me."
The company subsequently updated its privacy policy to hopefully remove any remaining nebulousness, as a follow-on Ars Technica writeup noted:
No facial mapping metrics, measurements, or other data used to predict demographics are stored.
For more coverage on this issue, see the following additional links:
- SceneTap Interview: San Francisco's Least Welcome Start-Up Explains Itself (Video) (AllThingsD)
- SceneTap cameras hit San Francisco bars, use facial recognition to find parties and privacy concerns (The Verge)
- Crowd-detection app SceneTap tries to allay privacy fears after rocky San Francisco launch (The Verge)
