While Meta may not be the first name that springs to mind in discussions about generative artificial intelligence, the company plays a significant role in the ongoing AI competition. With the introduction of its own model, Llama, and by integrating “Meta AI” into its prominent platforms—whether users are on board or not—Meta has firmly positioned itself within the AI landscape. The company even encourages exploration of personalized AI bots. Clearly, Meta is committed to its artificial intelligence initiative.

However, even for a company so deeply invested in AI, the recent developments come off as quite peculiar. Since 2023, Meta has been trialing user accounts generated through artificial intelligence on its platforms. The facades of these accounts, particularly their presence on Instagram, are currently trending, although they are also available on Facebook. Each account is verified and possesses a distinct personality, yet they are entirely fabricated, showcasing AI-generated imagery in their posts.

This scenario is unsettling and not entirely unprecedented—the accounts were established over a year ago but seem to have been mostly disregarded until now. Following a wave of online criticism, Meta has begun erasing their content from view.

The Unsettling Nature of Meta’s AI Users

The online community’s disdain for these artificial personas is easy to understand. For instance, consider “Liv” (username “himamaliv”), who claims to be a “proud Black queer momma of 2 & truth-teller.” Liv is not a real person, nor are the experiences she shares on her Instagram account genuine. Yet, the character persists, posting about empowering young girls, enjoying ice skating adventures with her family, and experiencing fun-filled days at the beach with “the kiddos.” Each narrative is accompanied by a corresponding image—assets that are, in reality, created by AI.

To its credit, Meta includes a watermark on each image, explicitly indicating its artificial origin; however, that does little to alleviate the creepiness of the scenario. Why would a generated “mother” be sharing fabricated beach outings involving nonexistent “children”? What genuine impact arose from a coat drive she takes pride in leading?

Credit: Jake Peterson

In an earlier post dated September 26, 2023, Liv mentions, “My backyard is my happy place… I’ve hosted countless birthday parties, cookouts, and girls’ nights here,” paired with a fabricated image of a picnic. The reality? Liv has done none of these things because, simply put, her existence is fictional. The life she claims to cherish is not real.

As of now, Liv follows 18 accounts, with 13 appearing to belong to similar AI-generated personalities. There’s Becca (dogloverbecca), who shares fabricated dog-centric content; Brian (hellograndpabiran), promoting himself as “everybody’s grandpa”; and Alvin the Alien (greetingsalvin), who, well, claims to be an alien.

Interestingly, not all content from these accounts is generated by AI. Some posts include videos that seem to portray real experiences, although AI-created videos can be incredibly realistic today. It’s believed that at least some of these videos, like a cooking segment featuring Carter, the AI dating coach, were authentic before Meta removed much of the associated material. Still, one cannot help but ponder who was behind these posts and why.

Viral AI Accounts: A New Twist on an Old Experiment

The peculiar aspect of this situation is that these accounts and their posts are not recent developments. Liv’s latest post dates back to March 8, 2024, which is characteristic of various AI-generated personas. For the most part, these profiles appear neglected, yet they still sport verification badges. As this article is being penned, Meta is actively scouring Instagram to delete these dubious accounts, while the Facebook versions seemingly continue to exist for now.

Credit: Jake Peterson

These accounts trace back to AI chatbots Meta unveiled in 2023, during a significant push into AI technologies. The focus at that time largely revolved around celebrity AI chatbots, allowing interactions with digital renditions of figures like “Tom Brady,” “Kendall Jenner,” and “Paris Hilton.” However, alongside these star-studded creations came lesser-known entities such as Liv, Brian, and Alvin. Engaging with them remains an option— simply visit Liv’s account for a chat. However, akin to most chatbot encounters, the dialogue is unlikely to be particularly enlightening.

It remains uncertain why these accounts are now attaining viral status, a year and a half post-launch. A reasonable assumption points to their sheer awkwardness. The profiles exude an unappealing vibe, layered with unsettling characteristics: the personalities crafted by Meta are off-putting (and potentially offensive); the posts themselves evoke discomfort (who genuinely wishes to view AI-generated images portraying fictitious children engaging in beach activities?); and the presence of verification undermines its original intent entirely.

In essence, while these accounts hail from 2023, they encapsulate a growing discontent towards the prevalence of artificial intelligence clumsiness as we continue into 2025. Reflections from Jason Koebler of 404media echo this sentiment:

The currently viral Meta AI profiles are old and are already inactive because they were such a colossal failure and were indistinguishable from AI spam. This ‘inevitable’ future Zuckerberg is trying to shove down our throats is being completely rejected.