The concept of smart glasses that can immediately recognise people may sound futuristic, but it has already sparked concerns over privacy. Experts say that the technology may alter the way people interact in public spaces but makes anonymity harder than ever.
Democratic senators are pushing Meta over its supposed plans to introduce real-time facial recognition technology in its Ray-Ban smart glasses. Senators Edward J. Markey, Jeff Merkley, and Ron Wyden, all Democrats from Massachusetts and Oregon, respectively, wrote to Meta CEO Mark Zuckerberg, stating concerns over the possible stalking, harassment, and selective intimidation that may result if the technology matches the faces it recognises with social media profiles without the person's consent.
The lawmakers want to know how Meta handles this type of information, how they get consent from bystanders, and whether they share any information. “Americans do not consent to biometric data collection simply by walking down a public street,” they wrote.
The company’s smart glasses are already under scrutiny when it comes to collecting user information. There were reports that contractors in Nairobi were viewing sensitive content collected by the smart glasses, such as intimate content. “In some videos, you can see someone going to the toilet or getting undressed,” one contractor stated.
Experts have expressed that the risks are increased with the integration of cameras with AI-based face recognition technology. Senior Policy Analyst John Davisson at the Electronic Privacy Information Center explained that the wearers cannot consent on behalf of everybody else and that the AI technology could exacerbate the data privacy concern.
The senators’ letter asked about the identification of the faces with Facebook or Instagram accounts, the consent mechanisms, and the storage or sharing of the biometric data. The senators have asked them to respond by April 6.
Meta explained that the videos could be filtered before being reviewed by humans and that the technology uses a combination of automated and manual processes to improve the technology.