The Georgetown Center on Privacy & Technology's Clare Garvie spoke at SXSW today on algorithmic bias and facial recognition technologies. The panel, "Facial Recognition: Please Search Responsibly" highlighted the Center's groundbreaking work on the widespread use of facial recognition technology--and the inherent problems it can present.
Speaking alongside Brian Brackeen, CEO and founder of Miami-based face recognition firm Kairos, and Arun Ross, professor in the Department of Computer Science and Engineering at Michigan State University, Garvie emphasized challenging questions about privacy, use vs. misuse, and the responsibilities of algorithm developers and companies to ensure privacy and civil liberties protections are built into their products
Brackeen discussed his company's efforts to address racial bias in facial recognition technology by building a more inclusive data set to train such programs. Garvie welcomed that approach, noting the importance of inclusive data to produce more refined and accurate results. The Privacy Center's 2016 report, "The Perpetual Line Up: Unregulated Police Face Recognition in America", revealed the widespread use of facial recognition technology by federal and state law enforcement, and highlighted that such technology may be less accurate on persons of color. The report noted the lack of any independent testing regime for racially-based error rates.
The Center's 2016 report was followed by a 2017 report on the use of face recognition technologies in U.S. airports, "Not Ready for Takeoff: Face Scans at Airport Departure Gates".
You can read more of the Privacy Center's outstanding work on facial recognition technology here, and follow their work in this space on Twitter at www.twitter.com/georgetownCPT.