US lawmakers have written to the social media company Meta demanding answers over proposals to introduce facial-recognition technology to its smart glasses, which the lawmakers say puts the privacy and civil liberties of citizens at risk.
Democratic senators have written to Meta’s chairman and CEO Mark Zuckerberg, warning that the “uniquely dangerous” plans could accelerate the “normalisation” of mass surveillance.
They warn that smart glasses could be used to capture images of thousands of people without their knowledge and instantly link them to names, workplaces or personal profiles, putting people at risk of stalking, harassment and targeted intimidation.
Meta’s plans, which were first disclosed by the New York Times, have raise particular concerns following the use of facial-recognition tools by US federal agencies to identify people involved in lawful protest, which is protected under the First Amendment.
“This abuse of facial-recognition tools demonstrates how easily real-time identification technologies can be repurposed to discourage political expression, target vulnerable communities and chill lawful dissent,” the letter stated.
Meta’s record over privacy questioned
Democratic senators Ed Markey for Massachusetts and Ron Wyden and Jeff Merkley for Oregan point to Meta’s poor record of protecting the privacy of its users, which – as Computer Weekly has previously reported – has frequently led to the company facing criticism.
According to the leaked memo, reported by the New York Times, Meta was advised to release the product “during a dynamic political environment where many civil society groups that would expect to attack us would have their resources focused on other concerns”.
The senators said that the disclosures suggested that Meta recognised the serious civil-liberties risks of facial recognition, and was planning to limit scrutiny by launching the technology when “the world is distracted by the Trump administration’s daily chaos”.
The senators point out that Meta abandoned facial recognition on its Facebook platform in 2011, citing wider ethical concerns about the use of facial-recognition technology. The technology was used to automatically tag people in pictures and videos when they opted in to the service. The company held internal discussions on introducing facial recognition on its smart glasses in the same year but decided not to go ahead.
“Five years later, Meta appears less worried about those societal concerns,” the senators wrote, “and is reportedly planning to deploy facial-recognition technology in one of the most dangerous possible settings: smart glasses.”
Concerns about linking facial recognition to profiles
The senators said they are particularly concerned that the integration of facial recognition with social media profiles would allow users to instantly access an individual’s name or social media profile.
“Such real-time identification would dramatically lower the barrier to doxxing and persistent tracking, enabling bad actors to connect a face in a workplace, name or online presence in seconds,” they said.
They warn that allowing strangers “frictionless access” to details about people’s identity raises serious risks of “stalking, harassment and targeted intimidation, particularly for women, LGBTQ+ individuals and other vulnerable communities.”
The senators are seeking answers from Meta over how the company would obtain proper consent from people who are captured by smart glasses, how long their biometric data would be stored and whether the data captured would be used to train machine learning or facial-recognition algorithms.
They are also seeking assurances from Meta about whether it plans to upload images of known individuals to create personalised databases for facial recognition, whether it plans to match images captured by smart glasses to profiles on Instagram and Facebook, and whether users of those platforms will have the option to opt out of having their data used in facial recognition.
The senators asked Meta to disclose what steps it intends to take to ensure facial-recognition systems do not disproportionately harm people of colour, immigrants, religious minorities, LGBTQ+ individuals or other vulnerable populations, and whether it intends to share information gathered by smart glasses with law enforcement agencies, including the Department of Homeland Security.