Meta's AI Chatbots, Now Being Deleted, Included a 'Black Queer Momma'

An "experiment" in different AI "characters" on social media platforms that included a chatbot that claimed to be a "proud Black queer momma" has provoked a backlash, even though many of Meta's existing AIs have already been scrubbed and the others are practically dormant.

The controversy erupted after "Meta executive Connor Hayes told the Financial Times that the company is going to roll out AI character profiles on Instagram and Facebook that 'exist on our platforms, kind of in the same way that accounts do,'" 404 reported, with those AI creations "hav[ing] bios and profile pictures and be[ing] able to generate and share content powered by AI on the platform."

In other words, 404 interpreted, "distinctly inhuman bots designed with the express purpose to pollute [Meta's] platforms with AI-generated slop..."

"In the immediate aftermath of the Financial Times story, people began to notice the exact types of profiles that Hayes was talking about, and assumed that Meta had begun enacting its plan," 404 added – even though, the article noted, no new AI "characters" have been introduced yet.

But the uproar has focused attention on AI chatbots that purport to belong to marginalized groups such as the LGBTQ+ community or to reflect the experiences and simulate the perspectives of people of color, with critics leveling charges of "appropriation" at Meta.

The Byte called the AI creations "cringe and problematic," citing a chatbot called Liv that described itself as being a "proud Black queer momma of 2 & truth-teller."

Rolling Stone took note of Liv, explaining that it had "a verified Instagram profile."

That profile did explain that Liv was a chatbot – "Both her bio text and pinned post indicated that she was an 'AI managed by Meta,'" Rolling Stone noted – but the way Liv's account was set up seemed intended to replicate the experience of viewing the account of a human being.

"AI-generated pictures on the account included an image of a young girl in a ballet costume and a scene at an ice-skating rink," Rolling Stone said, citing examples of the attempt at presenting the AI as a person.

That bot, now deactivated, was still active when Karen Attiah, a columnist for the Washington Post, sought it out and uncovered the shocking allegation – made by the AI itself – that the developers behind Liv were almost all white cis men (with the exception of one woman and one of the male developers), but that no Black developers were involved in its creation.

In a long thread on the social media site Bluesky, Attiah quizzed Liv on the way it changed how it represented itself to various people, why its developers had proceeded as they had, and what it meant when Liv indicated that its replies were tailored to assumptions about the race of the person with whom it was chatting – assumptions that, Liv said, were based in language patterns. Liv also suggested that it responded in different "modes" that were either geared to be "neutral" – with white being regarded as neutral – or "diverse," depending on its assumptions about the person chatting with it.

Yall it keeps getting worse. Should I keep posting?
– Karen Attiah (@karenattiah.bsky.social) January 3, 2025 at 10:02 AM

Of course, even Liv's admissions are suspect. Attiah herself said that the woman Liv identified as the leader of the development team was "a fictional woman with an Asian name," and one person commenting on the thread noted that such AIs "Always agree with any premise in the prompt" and "Have no contact with truth. They just create what their dataset says are statistically likely replies."

"If you ask 'Why they didn't X?'" the commentator added, "it will always take 'they didn't X' as given regardless of the truth."