Citing ‘Societal Concerns’, Facebook Dials Back Facial Recognition, Looks to On-device Approach

Amid intensifying scrutiny over the Facebook platform’s societal impact, the company has announced that it will discontinue its use of facial recognition.

The news came by way of a blog post this week by Jerome Pesenti, Facebook’s VP of Artificial Intelligence, who explained that while the company recognizes clear benefits in facial recognition technology, these are currently outweighed by “concerns about the place of facial recognition technology in society,” as well as an uncertain regulatory environment.

Decade-long Investment

As The New York Times reports, Facebook first introduced its facial recognition technology on the social media platform in late 2010, via a photo tagging feature designed to automatically identify the faces of individuals in images uploaded by users. The feature was designed to spur engagement, with notifications alerting end users when their friends and acquaintances had posted images in which they are featured on their profiles.

Beyond engagement, the photo tagging feature has also served a useful purpose for some in providing alerts when photos of them are uploaded without consent.

Facebook also integrated its facial recognition technology into its Automatic Alt Text service, which is designed to provide descriptions of images for the blind and visually impaired. As Facebook winds down its use of facial recognition in the coming weeks, the AAT service will continue to operate without functionality enabling it to attach names to the individuals pictured in images.

These applications are presumably the result of years of research and development on Facebook’s part, and the company’s enthusiasm for the technology was signaled earlier this year in a BuzzFeed News report that claimed the company was investigating the possibility of implementing facial recognition in a planned smart glasses product.

But, as the report noted, Facebook was also weighing the legal and privacy implications of doing so.

Is This About Facial Recognition, or Facebook?

Such concerns were understandable given the intensifying media scrutiny over facial recognition technology in recent years, especially with respect to the use of the technology by law enforcement authorities in the United States and repressive governments elsewhere. But they are also understandable in light of Facebook’s own troubles.

Last year, the company reached a $650 million settlement in a privacy rights lawsuit that had been working its way through the legal system for years. An initial suit was brought against the company in 2015 with a claim that Facebook’s face tagging feature violated Illinois’s Biometric Information Privacy Act (BIPA), and it later snowballed into a class action suit that was settled for $550 million before the presiding judge asked Facebook to return with a higher offer that would be more commensurate with the nature of the BIPA violation and the number of potential claimants.

And, as The New York Times notes, Facebook’s facial recognition technology was also a factor in a $5 billion fine issued by the Federal Trade Commission in 2019 over privacy complaints.

Facebook’s decision to limit its use of facial recognition arrives in the wake of those difficulties, and amid others. Former employee Frances Haugen recently became a high-profile whistleblower who testified to government officials about Facebook’s negligence over the use of its platform to organize violent movements including the January 6 insurrection, and about the baleful effects of its engagement-promoting algorithms on the mental health of users. Facebook’s parent company has responded by changing its name to Meta, with CEO Mark Zuckerberg seeking to draw attention toward the company’s efforts to champion virtual environments that he collectively describes as “the metaverse”.

The Baby in the Bathwater

However one interprets the rationale underlying Facebook’s retreat from facial recognition, the company is insistent on the technology’s clearest benefits, and is taking a nuanced approach that will not abandon the technology entirely. The company will not dispose of its DeepFace biometric algorithm software, and will continue to use facial recognition to help individuals to gain access to a locked account, or to authenticate when using financial services products or unlocking a device.

“These are places where facial recognition is both broadly valuable to people and socially acceptable, when deployed with care,” explained Facebook’s Pesenti in his blog post. “While we will continue working on use cases like these, we will ensure people have transparency and control over whether they are automatically recognized.”

Pesenti also went out of his way to highlight the “particularly valuable” use of facial recognition when used “on-device” in order to authenticate a given user without sending biometric data to an external server. “We believe this has the potential to enable positive use cases in the future that maintain privacy, control and transparency, and it’s an approach we’ll continue to explore as we consider how our future computing platforms and devices can best serve people’s needs.”

That could bode well for proponents of the on-device approach to biometric authentication such as those affiliated with the FIDO Alliance. And while it may irk those who disagree with the on-device ethos, the broader support for facial recognition expressed by Facebook’s head of AI – even as the company scales down its use – may offer a welcome contribution to the ongoing discourse around the ethical use of AI and biometrics.

Sources: Meta, The New York Times

–

November 2, 2021 – by Alex Perala