Policy@Intel
A place to exchange ideas and perspectives, promoting a thriving innovation economy through public policy
643 Discussions

Building Privacy into Facial Detection Technology

Intel_Blog_Admin
Employee
0 0 45
 

brianhuseman-thumb-150x208-thumb-150x208-thumb-150x208By Brian Huseman, Senior Policy Counsel

Today, I had the opportunity to participate in the Federal Trade Commission’s workshop on facial recognition, facial detection, and digital signage technologies, called “Face Facts.”As usual, the FTC staff did a tremendous job of pulling together a great series of panels, which explored in depth a particularly timely set of new technologies.

I spoke about Intel’s facial detection software called AIM Suite (Audience Impression Metrics), which uses Anonymous Viewer Analytics (or AVA), a technology that through pattern detection algorithms can detect faces and identify a person’s gender and age range.Facial detection technology is being used in digital signs, which allow advertisers to display ads that are more relevant for consumers. Consumers also get more interactive retail displays (such as those being provided by Adidas, Kraft, and Procter & Gamble) with the potential for real-time coupons and QR code discounts.

Shoe AIM pic

Intel strongly believes in the principle of “Privacy by Design,” as first championed by the Ontario, Privacy Commissioner Dr. Ann Cavoukian, and we have built privacy protections into our facial detection technology.For example, Intel’s AIM Suite software does not record or capture personally identifiable information. Facial representations are detected from optical sensors and processed to collect anonymous data. No individually identifiable data is collected or categorized. No image is recorded or captured.Fred Carter from Dr. Cavoukian’s office was a fellow panelist at the workshop and discussed the recent white paper issued by their office, which highlighted Intel’s AIM Suite as a good example of the concept of Privacy by Design.

Harley Geiger from the Center for Democracy and Technology (CDT) also was a fellow panelist today.In advance of the workshop, CDT released a must-read paper on the privacy implications of facial recognition and detection technology. CDT has set forth varying levels of consumer privacy interests based upon the level of identification of the individual (from just counting individuals to targeting them with content up to the identification of individuals via facial recognition).CDT also spoke about the Digital Signage Federation self-regulatory privacy guidance which has done a good job of shaping this nascent industry.Importantly, Intel agrees with CDT’s call for comprehensive federal privacy legislation dealing with all practices and technologies instead of regulation of facial recognition and detection alone.

Intel is pleased that the FTC is on the forefront of examining these emerging technologies.There are obvious privacy implications from facial detection and digital signage, but rather than limiting the use of the technologies themselves, strong privacy protections can be built into the technology.I am glad to have participated in the FTC event today and look forward to continued discussion of these issues.