Meta scraps its celebrity AI chatbot feature less than a year after launch and now expects people to use Meta's new AI Studio to customize their own AI avatars
“What's this?” — “Oh, ok.” — “Anyway....” — RE: https://www.threads.net/... X: Brendan Brady / @bradybrendan1 : I will miss you AI Paris Hilton! Brett Dashevsky / @brettdash_ : this was so predictable. just a splashy way for meta to show off their ai-tech and get user data @ttp_updates : A company that is struggling to stop its algorithms from recommending CSAM should not be wasting time on AI chatbots. Instead of investing in trust and safety, Meta chased AI hype - and flopped. Rachel Metz / @rachelmetz : This comes as no surprise to anyone who read the comments under the first posts from “Billie”, aka Kendall Jenner, that were filled with people who were confused and annoyed. Walter Hickey / @walthickey : you would think that if anyone could be replaced it's “famous for being famous” types, but with this flop I feel a crucial reason that AI can't replace human creativity is that it lacks, literally, “je ne sais quoi” https://www.theinformation.com/ ... Tim Baysinger / @tim_bays : feels like we're starting to see an idea of what people want/don't want from AI Kaya Yurieff / @kyurieff : New: After spending millions of dollars to license celebrities' likenesses for AI chatbots, Meta is scrapping the effort after it didn't take off with users: https://www.theinformation.com/ ... w/ @SylviaVarnham & @KalleyHuang Amir Efrati / @amir : breaking: celebrity AI bots are not a thing :) https://www.theinformation.com/ ... $meta @kyurieff [image] See also Mediagazer
Context & Ripple Effects
Meta’s consumer-facing character experiments emerged alongside a broader AI strategy that had emphasized improving Reels engagement and advertising performance with AI tools AI tools for views and ad targeting. The celebrity bots also fit the then-growing idea of synthetic companions as a new consumer-internet category synthetic companions as a consumer product.
The pivot from a curated celebrity format to user customization is consequential because it changes both the source and scale of characters on Meta’s services. Later coverage of Meta’s plans for AI characters across its apps AI characters spreading across Meta’s apps shows this was less a retreat from social AI than a change in how it would be supplied.
First-order effects
- People using the celebrity characters lose that product, while Meta redirects the consumer AI proposition toward AI Studio and user-made avatars.
- Meta replaces a limited, centrally designed roster with a creation tool, making customization—not celebrity branding—the immediate product focus.
Second-order effects
- A user-generated model can expand the variety of characters without Meta commissioning each one, but it also makes safeguards around misleading or inappropriate profiles more operationally important; Meta later removed viral AI profiles after backlash backlash over viral AI-generated profiles.
- The shift tests whether Meta’s distribution can make avatar creation a repeat-use social feature rather than a novelty driven by recognizable personalities.
Third-order effects
- If Meta continues to populate its apps with user-created AI characters, the company’s advantage may increasingly rest on distribution and moderation systems rather than a small set of branded bots.
- The model also raises a durable trust trade-off: more creation and sharing surfaces can increase exposure to accidental disclosure, a risk later visible in Meta AI’s public feed sensitive chats appearing in a public feed.
The trend: Social platforms are moving from celebrity-led AI demonstrations toward scalable, user-generated AI identities distributed inside existing social networks.