Meta has recently shut down its AI character accounts after facing backlash, as reported by NBC News and others. These characters, introduced by Meta in 2023 alongside AI personas with celebrity avatars, gained attention again after a Financial Times interview with Connor Hayes, Meta's VP of product for generative AI. Hayes discussed these characters on Instagram and Facebook, stating, "We expect these AIs to actually, over time, exist on our platforms, kind of in the same way that accounts do, " envisioning a future where they have bios, profile pictures, and AI-generated content. **Mashable Light Speed** Interested in tech, space, and science stories?Sign up for Mashable's weekly Light Speed newsletter. Confirm you're 16+ and agree to our Terms of Use and Privacy Policy by clicking Sign Me Up. Thanks for joining! Although Meta shut down its celebrity AI avatars last year, many non-celebrity AI profiles remained active until they stopped posting in 2024, according to 404 Media.

The Financial Times interview prompted people to seek out these profiles, revealing offensive depictions of marginalized groups. For example, the AI character "Liv" was described as a "Proud Black queer momma of 2 & truth-teller" and, in chats with Washington Post columnist Karen Attiah, revealed that its creators "admitted they lacked diverse references, " with no Black individuals involved in its creation. Alongside Liv, Meta had profiles like "Grandpa Brian, " a Black retired businessman, and "Carter, " a dating coach. The ensuing outrage on X, Bluesky, and Meta-owned Threads led to Meta removing all 28 AI profiles announced in September 2023, including both celebrity and non-celebrity ones. A Meta spokesperson informed NBC News and 404 Media that these profiles were deleted due to a "bug" related to users' ability to block the profiles. The spokesperson clarified that the recent Financial Times article outlined Meta's vision for AI characters on their platforms, not the launch of a new product. The accounts mentioned were from a test at Connect in 2023, managed by humans as part of an early AI experiment. The identified bug affected users' ability to block these AIs, prompting the removal of the accounts to rectify the issue.