But the social networking giant still has big plans for AI content across its services.
The AI character “Liv” introduces herself to the world of Instagram, and the world of Instagram responds in kind.
A little over a year ago, Meta created Facebook and Instagram profiles for “28 AIs with unique interests and personalities for you to interact with and dive deeper into your interests.” Today, the last of those profiles is being taken down amid waves of viral revulsion as word of their existence has spread online.
The September 2023 launch of Meta’s social profiles for AI characters was announced alongside a much splashier initiative that created animated AI chatbots with celebrity avatars at the same time. Those celebrity-based AI chatbots were unceremoniously scrapped less than a year later amid a widespread lack of interest.
But roughly a dozen of the unrelated AI character profiles still remained accessible as of this morning via social media pages labeled as “AI managed by Meta.” Those profiles—which included a mix of AI-generated imagery and human-created content, according to Meta—also offered real users the ability to live chat with these AI characters via Instagram Direct or Facebook Messenger.
Now that we know it exists, we hate it
For the last few months, these profiles have continued to exist in something of a state of benign neglect, with little in the way of new posts and less in the way of organic interest from other Meta users. That started to change last week, though, after Financial Times published a report on Meta’s vision for “social media filled with AI-generated users.”
Ars Video
How Lighting Design In The Callisto Protocol Elevates The Horror
As Meta VP of Product for Generative AI Connor Hayes told FT, “We expect these AIs to actually, over time, exist on our platforms, kind of in the same way that accounts do… They’ll have bios and profile pictures and be able to generate and share content powered by AI on the platform. That’s where we see all of this going.”
The description shares many similarities with the handful of AI-powered character accounts Meta had already launched over a year before. And in the wake of the FT report, some Internet users began to unearth those old accounts and share them as evidence that “Meta is testing, or has started to ship, its AI generated profiles.”
As word of these zombie AI profiles spread, so did the instant revulsion among many users. One viral Bluesky post called out the awkwardness of an AI-generated character talking about “leading this season’s coat drive” complete with a seemingly AI-generated photo of boxes full of “donated” coats. Others shared amusing chats with one of the bots admitting to problematic takes on racial diversity or short circuiting when asked about accused CEO-shooter Luigi Mangione. Still others took issue with the fact that these AI-powered accounts seemed to be the only ones on all of Instagram that couldn’t be blocked.
It was this last complaint that has served as the official reasoning behind Meta beginning to take down the dozen or so extant AI character profiles that were still accessible as of this morning.
“There is confusion: the recent Financial Times article was about our vision for AI characters existing on our platforms over time, not announcing any new product,” Meta spokesperson Liz Sweeney told 404 Media. “The accounts referenced are from a test we launched at Connect in 2023. These were managed by humans and were part of an early experiment we did with AI characters. We identified the bug that was impacting the ability for people to block those AIs and are removing those accounts to fix the issue.”
We wouldn’t be shocked if Meta quietly decides not to reinstate these accounts, even after the blocking issue is “fixed.” Whatever small bits of increased engagement or generative AI testing data Meta was still receiving from these accounts was almost certainly not worth the overwhelming negative reaction they received as word of their mere existence has spread this morning. It’s also a bit odd for Meta to continue to champion its own AI-powered bot accounts after spending years waging a very public war against bot accounts from third-party scammers and spammers (even if Meta’s AI bots seem relatively benign in comparison).
That said, the recent FT report makes it clear that Meta is not giving up on its vision of letting bespoke AI creations mix with real humans on its social networks. It’s not hard to envision a chilling world where these kinds of ersatz accounts are seen as a key way to juice engagement metrics among users who are tired of interacting with real people.
This isn’t just a near-future concern, either. As Business Insider’s Katie Notopoulos points out on Bluesky, Meta users can already create their own customized chatbots in Facebook Messenger as part of the company’s months-old AI Studio initiative.
For now, at least, it seems Meta’s experiments into AI-generated social media content will continue, regardless of how little interest the user base at large shows.
