Facial recognition research erases trans and nonbinary people

A February 2019 study of facial recognition systems found that these systems are not designed with transgender and nonbinary people in mind. Researcher Os Keyes studied 30 years of facial recognition research, including 58 separate research papers, and found that more than 90% of the time researchers followed a binary gender model, more than 70% of the time viewed gender as immutable, and, in gender-specific research, more than 80% of the time viewed gender as a purely physiological construct. Overall, Keyes concluded, automated gender recognition systems fundamentally ignore the existence of transgender people - with important consequences as automated systems to send alerts or grant access to spaces such as airports and bathrooms are developed and deployed that depend on facial recognition or automated gender recognition.


Writer: Matthew Gault

Publication: Motherboard

Publication date: 2019-02-19