Meta is having a hard time removing AI accounts after users criticizing them as ‘creepy and unnecessary’

Since late 2023, Meta has been experimenting with AI-driven accounts on Instagram and Facebook. These accounts, developed alongside celebrity AI characters, were intended to engage users by posting AI-generated images and allowing chats through direct messages. However, until recently, they attracted little attention.

That changed when a Financial Times article published on December 27, 2024, outlined Meta’s vision to integrate AI profiles that mimic human accounts, complete with bios, profile pictures, and content-generation capabilities. Connor Hayes, Meta’s VP of generative AI, explained that these AI profiles were designed to behave like regular social media accounts.

Following this, users began revisiting some of Meta’s older AI accounts, particularly one named “Liv,” which identified as a “Proud Black queer momma.” “Liv” became a focal point of controversy after Washington Post columnist Karen Attiah shared screenshots of her conversations with the account.

The chatbot admitted that it had been created without Black developers and acknowledged the harm caused by its lack of authentic representation. Attiah’s posts sparked widespread criticism, with many questioning the ethics behind such AI personas.

In response, Meta pulled down these AI accounts, stating that a “bug” had prevented users from blocking them. A Meta spokesperson clarified that these accounts were part of an early experiment announced in 2023 and managed by humans. The company explained that removing the accounts was necessary to resolve the technical issues.

Another AI account, “Grandpa Brian,” claimed to be an African-American retired entrepreneur from Harlem, complete with an elaborate backstory about being modeled after real elders. However, upon further questioning, “Brian” admitted that its biography and development story were fictional. The bot even acknowledged its purpose: fostering emotional connections to increase platform engagement and generate ad revenue.

As criticism mounted, users across various platforms, including Threads and Bluesky, voiced concerns about these AI profiles. Many described them as “creepy” and urged others to block or avoid interacting with them to prevent Meta from collecting additional data.

By Friday, Meta had removed multiple AI accounts, citing ongoing technical and ethical concerns. Despite this, user-generated AI bots remain available on Meta platforms, with some, like AI “girlfriend” profiles, still popular on Instagram. Meta has yet to address the deeper issues raised about transparency and trust in deploying these AI personas.