Facial recognition technology, whether you’re a friend or foe, has a robust future, something clearly on display during the National Telecommunications & Information Administration’s (NTIA) March 25 multi-stakeholder meeting on creating a code of conduct for the technology. With the NTIA still in the gathering and learning stage, the meeting featured presentations from technology experts, some industry representatives and the Federal Trade Commission (FTC).
Whether for authentication, identification or other uses, the technology’s complex integration into daily business—including banking, healthcare, transportation and social media—and rapid evolution pose a challenge for stakeholders in coming up with a focused and agreeable code of conduct. Carried over from the last meeting, stakeholders submitted use cases to help flesh out where the concerns and benefits reside for crafting a code.
Carnegie Mellon researcher Alessandro Acquisti led off the proceedings discussing his work in the field and the potential linkages made possible by facial recognition between the online and offline worlds. IBG’s Olga Raskin presented on the processing abilities of facial recognition, including detection, cropping, recognition, grouping and tagging.
“Online facial processing,” she noted in closing, “is improving and impressive.”
Attendees also heard from identification assurance and management solutions firm Paycasso, a service that provides advanced one-to-one photo matching for a number of companies including banks, healthcare providers and airports. Paycasso CEO Russell S. King noted the technology provides beneficial uses for detecting and preventing fraud and identity theft.
FTC Staff Attorney Amanda Koulousias also went over the agency’s facial recognition report, culled from a roundtable held in 2011, a number of multi-stakeholder comments and underpinned by its 2012 staff report. The FTC was clear that it is not trying to affect the NTIA proceedings but did express hope the multi-stakeholder process will eventually provide an enforceable code for the FTC under the deceptive prong of Section 5.
On Wednesday, stakeholders were also asked to call in for a conversation with Gwendal Le Grand, head of the CNIL’s IT Experts Department, to discuss the Article 29 Working Party’s work on facial recognition technology.
Moving forward, stakeholders were asked to look at the submitted use cases to help develop priorities, or the “low-hanging fruit,” that everyone can agree on to start constructing a code. The NTIA’s John Verdi organized the use cases into three main categories: facial recognition technology used for identification by organizations, facial recognition technology used by individuals for identification and facial recognition technology for authentication. He also included a fourth category for the more idiomatic cases.
Stakeholders seemed to have difficulty agreeing on how best to go about prioritizing and collectively analyzing the use cases, but most agreed that a short three- to five-minute presentation on a use case by each willing stakeholder would be helpful moving forward.
The Center for Digital Democracy’s Jeffrey Chester reignited calls for Google and Facebook to appear at the next meeting to present their use of facial recognition. Both declined calls to present at Tuesday’s meeting.
With a quick turnaround and a load of use cases to go over, the group decided to cancel the next meeting, originally scheduled for April 8, and instead convene on April 29.