Meta Platforms is currently testing facial recognition to combat scam ads that misuse celebrity images. The company’s aim is to reduce the number of fraudulent advertisements known as “celeb-bait,” where public figures’ faces are used without permission. These ads trick users into clicking on malicious links.
Meta’s system compares faces from flagged ads to the profile pictures of public figures stored on Facebook or Instagram. If it finds a match and determines that the ad is fake, it gets blocked before anyone sees it. This automated process is designed to work alongside existing machine learning systems, which Meta already uses to scan ads for policy violations. Early testing with a small group of celebrities has shown promising results, according to the company.
New Facial Recognition Tool Aims at Account Recovery Too
Along with flagging scams, Meta is testing another use for facial recognition technology: helping users regain access to their hacked accounts. Currently, if someone gets locked out of their account, they can verify their identity by submitting a government ID. However, this can be time-consuming. Meta’s new solution would let users upload a video selfie, which the system compares to the account’s profile picture to verify identity more quickly.
Though the technology could provide a faster alternative to the existing process, Meta’s use of facial recognition has already sparked privacy concerns. In 2021, the company discontinued its auto-tagging facial recognition system over fears about data misuse. Now, as Meta brings facial recognition back for specific tasks, it says it will delete facial data once the comparison is complete.
We’re testing new ways to make it harder for scammers to bait people and easier for people to regain access to their accounts using facial recognition technology.https://t.co/eLii9xqDFg pic.twitter.com/5sgFcJJ7xV
— Meta Newsroom (@MetaNewsroom) October 21, 2024
Meta Smart Glasses Highlight Real-Time Privacy Risks
On a different front, the conversation about facial recognition tech has extended to wearables. Earlier in October, two Harvard students demonstrated how Meta’s Ray-Ban smart glasses, paired with facial recognition software, could be used to track personal details in real time. Their system, called I-XRAY, allowed the students to gather private information such as names and addresses by scanning faces and matching them with public databases.
What’s alarming about this is that the tech required to pull off this kind of surveillance is already widely available. Using off-the-shelf technology, the students were able to expose the potential dangers that come with embedding facial recognition in everyday products like glasses. Meta’s Ray-Ban glasses do come with a small light to indicate when recording is happening, but in a crowded space, this light can easily go unnoticed.
AI Misidentifications Fuel Privacy Debates
Meta’s facial recognition tests land in the middle of a broader conversation about the dangers of relying on AI for identifying people. Earlier this year, in May, a facial recognition system wrongly accused a British shopper of stealing at a Home Bargains store. The technology flagged her as a repeat offender, leading to her being escorted out and banned. The company responsible for the tech, Facewatch, later admitted it made a mistake.
This wasn’t an isolated case, as several incidents have raised concerns about the accuracy and fairness of AI systems. In 2023, a Black teenager was wrongly identified by a skating rink’s facial recognition cameras as someone who had been previously banned. There’s been a string of other stories like this, all pointing to the fact that AI, as powerful as it can be, is far from infallible.
Facial Recognition in Sports Venues Sparks Protests
Not just in retail or online, facial recognition technology is also starting to become a fixture in public spaces like sports venues. Back in August, a group of protestors gathered outside Citi Field, home of the New York Mets, to express their concerns over its use in MLB and NFL stadiums. These arenas have been introducing systems like Go-Ahead Entry, which allows fans to bypass security lines by uploading a selfie for faster access.
While this sounds like a time-saver, privacy advocates argue that it opens up a new avenue for surveillance. Critics warn that the data collected through these systems could easily be accessed by law enforcement or other third parties without sufficient oversight. Groups like Amnesty International and Fight for the Future have been educating fans about the risks associated with these systems and demanding greater transparency from the venues and tech providers.
Facial Recognition: Public Distrust Grows, Regulation Lags Behind
Meta’s renewed interest in facial recognition comes at a time when governments are starting to take action. In May of this year, Microsoft updated its terms of service for Azure OpenAI, banning police departments in the U.S. from using its AI tools for facial recognition purposes. The move was part of a growing push to limit the technology’s use in law enforcement due to the risks of misidentification and data misuse.
In New York, lawmakers are currently considering regulations that would ban the use of facial recognition in public settings, including residential buildings. Two bills are on the table, aimed at controlling how facial recognition can be used for identifying individuals in public spaces, reflecting increasing unease about its unchecked spread.
While regulations slowly catch up, Meta and other tech giants continue to develop new ways to implement facial recognition technology. The balance between innovation and privacy remains a sticking point, and how companies like Meta navigate these challenges will determine the future of facial recognition tech in both public and private sectors.
Last Updated on November 7, 2024 2:24 pm CET
