Just days after Meta executives publicly attacked reports that the company had embedded facial recognition technology in its smart glasses app, the code in question is no longer there.

Last week, Wired and independent security researcher Buchodi reported that Meta’s companion app for Ray-Ban and Oakley smart glasses contained a complete but inactive facial recognition system known internally as NameTag.

Buchodi verified the existence of face detection models, biometric matching systems, local databases, and notification tools capable of identifying people viewed through the glasses.

Meta strongly disputed the reporting, and company executives took to social media to argue that the feature did not exist as described in the report.

Meta’s communications head Andy Stone went so far as to call the coverage “intellectually dishonest” and “pure advocacy-driven clickbait.”

Meta’s Chief Technology Officer Andrew Bosworth – or Boz, as most people know him – took to X to say how disappointed he was with the publication.

Code libraries are no longer there

But on Friday, one day after the investigation was first published and as it began to gain traction, Meta released an updated version of its Meta AI app that removed the facial recognition components identified by Wired and the researchers.

According to Wired’s analysis, code libraries related purely to face recognition are no longer present in the latest build.

Meta has not explained whether the removal was planned before the reporting appeared or whether it was a direct response to the scrutiny.

The company has also not said whether the technology could return in a future release.

It’s likely to raise more questions about the original findings, such as why biometric recognition infrastructure was embedded in software distributed to tens of millions of devices before public disclosure.

Engineering leader and privacy advocate Richard Kersey told Cybernews that this move amounted to “a race to embed security architecture before the legislators catch up.”

“They [Meta] have billions of users and a near-zero marginal cost to collect biometric data – not because it's the right architecture, but because they can,” Kersey said.

As Buchodi wrote in his original analysis, the issue wasn't just Meta secretly identifying people but that the "complete apparatus to do exactly that was already present on user devices."

---

Unlock more exclusive Cybernews content on YouTube.