Meta is further backing away from facial recognition. The company has announced it's shutting down the Face Recognition system on Facebook in the "coming weeks." You'll no longer be automatically recognized in photos if you opted in to the feature, and the Facebook team will delete over a billion facial recognition templates. Automatic Alt Text's descriptions for the visually impaired will also stop naming people detected in photos.
The company said it made the decision following "growing concerns" about the broader use of facial recognition, including "uncertainty" about regulation of the technology. While Meta believed face recognition could still be helpful in some situations, such as gaining access to a locked account, it felt a "narrower," more privacy-oriented approach was a better fit than the broader outgoing strategy. On-device recognition wouldn't require sharing data with outside servers, for example.
There was already significant pressure to shut down the system. Meta (then Facebook) settled a privacy lawsuit over facial recognition for $650 million in 2020, and it likely isn't eager to grapple with similar cases. Governments ranging from Massachusetts to the European Union have either passed or are strongly considering at least partial bans on facial recognition use. Facebook's opt-in system may not have explicitly run afoul of those laws, but the message is clear: wide-ranging facial recognition systems aren't very welcome.
As it stands, Meta isn't alone. Amazon now has an indefinite ban on police use of its Rekognition platform, and Google has declined to sell facial recognition tech since 2018. Although Meta hasn't completely given up on the concept, it's clearly eager to be seen as forward-thinking — and continue mending its less-than-stellar reputation for privacy.