The philosophical debate sparked by Meta’s “NameTag” feature touches on the very nature of technological existence: if code is sitting on millions of devices, but the public hasn’t been given the “on” switch, does that feature actually exist? For the engineers and executives at Meta, the answer appears to be a convenient “no.” When journalists at WIRED discovered that robust, functional code for a facial recognition system had been lurking inside the Meta AI app—which is installed on tens of millions of smartphones—the company’s response was one of semantic maneuvering. They insisted that because the feature hadn’t been launched to consumers, it didn’t really exist. This triggered a strange public standoff where the company’s spokesperson claimed the feature wasn’t there, even as independent researchers confirmed that the code was fully capable of identifying faces.
The controversy highlights a common friction between tech companies and the public regarding transparency. WIRED revealed that the NameTag code had been quietly sitting in the Meta AI app since at least the beginning of the year, growing in complexity as the months passed. By May, the system was so far along that an independent researcher was able to use the hidden code to successfully identify a photograph of the philosopher Michel Foucault. This discovery makes Meta’s insistence that the feature was non-existent feel somewhat hollow, if not entirely disingenuous. Shortly after WIRED exposed the presence of the code, Meta promptly removed it from the application, an action that subtly acknowledged the code’s reality far more than their official denials ever did.
Adding to the confusion, the company’s internal messaging has been wildly inconsistent. While communications staff were busy telling the press that NameTag didn’t exist, Meta’s Chief Technology Officer, Andrew “Boz” Bosworth, appeared on a podcast and provided a remarkably detailed breakdown of how the feature is intended to work. During the interview, he explicitly described a system where glasses-wearers could identify people they had previously met, framing this as a specific, functional capability. When confronted with this contradiction, Meta pivoted to a semantic defense, emphasizing that Bosworth used the word “would” rather than “is,” suggesting he was merely discussing a hypothetical concept. To the outside observer, however, hearing a high-ranking executive explain the mechanics of a system while simultaneously claiming the system doesn’t exist feels like a masterclass in corporate gaslighting.
Meta’s defense rests on an important distinction: they claim they are not building a tool to connect glasses to a giant public database of strangers. Instead, they argue that NameTag is intended to be a benign helper for those with visual impairments—a way to identify people you have already met or need to remember. By framing it this way, the company attempts to gain the moral high ground, suggesting that their critics are misinterpreting a tool designed for accessibility as an engine for mass surveillance. While this is a plausible and perhaps noble use case, it doesn’t resolve the fundamental issue of the company shipping hidden, high-stakes technology to millions of users without their explicit knowledge or consent.
When broken down, this situation reveals a widening gap between how tech giants deploy software and how the public expects them to act. In the modern era, software is no longer a static product that is finished, shipped, and then left alone; it is a living, breathing entity that changes constantly on our devices. Meta’s approach seems to be to build, test, and host potentially controversial features in the background of apps that millions of people already use, keeping them dormant until they determine the political or regulatory climate is right for a public rollout. From a development standpoint, this is highly efficient; from a trust standpoint, it is incredibly damaging. It treats the user’s smartphone less like a personal tool and more like a testing ground for experimental features that they haven’t yet agreed to host.
Ultimately, whether NameTag “exists” is a question that depends entirely on whether you prioritize the company’s internal branding or the technical reality of the software residing on your phone. Even if we accept Meta’s claim that this was merely an “exploration,” we have to acknowledge that the company possessed a functional facial recognition tool that sat ready for deployment on millions of handsets. By arguing over the definition of the word “exist,” Meta is choosing to prioritize legal and rhetorical safety over the transparency that users deserve. In a world where devices are increasingly capable of surveilling our environments, “existential” debates about code won’t satisfy a public that is becoming increasingly wary of how their personal data and privacy are being managed behind the scenes. Moving forward, the conversation shouldn’t be about whether code can exist, but whether it should occupy our space without our permission.
