AnyVision is calling on the National Institute of Standards and Technology (NIST) to draft guidelines for the responsible use of facial recognition technology.
The company made its request in the form of an open letter, which was framed as a response to NIST’s new draft publication, Artificial Intelligence and User Trust (NISTIR 8332). The draft publication is intended to stimulate discussion about how humans trust artificial intelligence systems, and whether such trust can be quantified and measured.
AnyVision’s response was penned by the company’s CEO, Avi Golan, who praised NIST for its development of “cogent guidelines” with respect to trust and AI. He went on to suggest that NIST extend this effort to the realm of facial recognition and video surveillance specifically, taking particular note of the threat of racial bias that researchers have documented among some biometric algorithms.
“AI companies must continue to purge demographic bias from their algorithms and be transparent about their methodology and the training data used to develop their models,” he wrote.
Golan also pointed out that European Commission officials have recently taken “steps in the right direction” by drafting safeguards on how AI can be used and what levels of risk are appropriate for different use cases – for example, in credit scoring versus law enforcement. Golan framed this as a “level of nuance” that is unfortunately “missing from most discussions today related to facial recognition.”
That lack of nuance has directly impact AnyVision in the past, with minority investor Microsoft having divested from the company over media reports concerning the use of its technology to surveil Palestinians in the West Bank – reports for which an external auditor ultimately found no evidence.
AnyVision previously drew attention to the need for ethical guidelines for the development of facial recognition technology at a conference hosted by Fordham University Law School earlier this year. Its own technology recently delivered a strong performance in NIST’s 1:1 Face Recognition Vendor Test.
June 3, 2021 – by Alex Perala