A U.S. Government Accountability Office report has urged Congress to develop an overarching federal framework for facial recognition. The report confirms that more businesses are using the technology than when the last report came out in 2015. Concerns abound when it comes to privacy, ethics, and consent, which is why a framework is needed.
A summary of findings:
- Some companies have started using facial recognition to monitor the spread of COVID-19.
- From 2016 to 2019, the global facial recognition technology market generated $3b to $5b in revenue. From 2022 through 2024, that's expected to grow to $7b to $10b.
- The number of participants in NIST’s Face Recognition Vendor Tests rose from 16 in 2013 to 99 in 2019, which suggests that more companies are developing facial recognition algorithms.
- Facial recognition technology is often dependent on the large-scale collection of facial images.
- Concerns include the inability of individuals to remain anonymous in public or the use of the technology without people's consent.
- Although the accuracy has increased dramatically in recent years, differences in performance exist for various demographic groups. The technology does better, generally, on lighter-skin men and worse on darker-skin women. It doesn't perform well on the elderly and children.
- New York City will reassess the NYPD's use of facial recognition technology after police used it to investigate a Black Lives Matter activist.
- Police in Miami, Fla., also used facial recognition from the firm Clearview AI to identify a woman accused or throwing rocks at officers during a May protest.
- A London court ruled that it's unlawful for South Wales Police to use automatic facial recognition to surveil crowds, which sets a precedent for how U.K. police can use the technology.