We’ve covered facial recognition in the 4Cast before, but this week we’re taking another look, after Amazon’s facial recognition wrongly identified 28 members of Congress as being criminals. This error is newsworthy on its own, but it also highlights some major problems and incorrect assumptions. For instance, the underlying databases that facial recognition uses often have implicit biases (see Can We Make Non-Racist Facial Recognition?)
Add to that additional news that is raising other issues and, to be honest, the creepiness factor. The biases built into this type of software is only one of the concerns.
- Schools can now get free facial recognition software to track students [Fast Company] “The tech is currently being tested at a school in Seattle (specifically, the school that the founder’s kids attend) where kids can unlock a gate by smiling at a surveillance camera—which may sound chilling to anyone who has read Nineteen Eighty-Four. It will reportedly be piloted in the state of Wyoming later this year.”
- Police accused of deploying facial recognition ‘by stealth’ in London [Independent]”Scotland Yard pledged that members of the public would be informed using leaflets and posters in Stratford, but monitors said people were not being properly warned before walking into the scanning zone.”
- Canadian Malls Secretly Tracked Shoppers’ Age, Gender Using Facial Recognition Technology [Gizmodo] “Canadian outlet CBC reports that the mall’s parent company, Cadillac Fairview, has admitted to using the software without shopper consent, but the company claims they only collect limited data. The software has been in use since June. “
- AI system looks at faces and rates them for responsibility, happiness, aggression, attractiveness, weirdness, and emotional stability [Boing Boing] “The Biometric Mirror scans people’s faces and uses AI to compare their faces against a database of other faces to produce a personality profile of the scanned person. The reports include ratings for the individual’s responsibility, happiness, aggression, attractiveness, weirdness, and emotional stability.”
From the Ohio Web Library:
- Facial Recognition Technology: An Overview (Crelin, J. (2017). Facial Recognition Technology: An Overview. Points Of View: Facial Recognition Technology, 1. )
- Counterpoint: Facial Recognition Technology Should Be Regulated to Address Privacy Concerns (Counterpoint: Facial Recognition Technology Should Be Regulated to Address Privacy Concerns. (2017). Points of View: Facial Recognition Technology, 1. )
- Duping facial recognition technology (Julien Gignac Toronto, S. (2018, June 7). Duping facial recognition technology. Toronto Star (Canada). )