Meta Did Not Have Permission to Make Its Flirty Celebrity Chatbots—AI Responses Tweaked for Teens Ahead of Senate Probe

Photo Credit: Taylor Swift by Eva Rinaldi / CC by 2.0

Meta scrambles to change its flirty celebrity AI chatbots ahead of a Senate probe so they no longer generate inappropriate responses to underage users.

Facebook parent company Meta has removed some of its AI celebrity chatbots, with “temporary changes” to its chatbot policies related to teens. The changes come ahead of a Senate probe launched to investigate reports of the chatbots engaging in “sensual” conversations with minors.

Last week, Senator Josh Hawley (R-MO) said he was launching an investigation into Meta following a report from Reuters that the company was permitting its AI chatbots to engage in inappropriate conversations with teens and children.

While many of Meta’s chatbots were user-created using a tool for the purpose, Reuters found that at least one Meta employee had produced no less than three chatbots, including two based on megastar Taylor Swift. Other celebrity chatbots on Meta’s platforms include Scarlett Johansson, Anne Hathaway, and Selena Gomez, as well as teen film star Walker Scobell. All the virtual celebrity bots have been shared on Meta’s Facebook, Instagram, and WhatsApp platforms.

A Meta spokesperson told Reuters that the company’s AI tools should not have created intimate images or lewd conversations. In particular, they blamed Meta’s production of images of scantily clad female celebrities on “failed enforcement of its own policies,” which prohibit such content from being produced.

“Like others, we permit the generation of images containing public figures, but our policies are intended to prohibit nude, intimate, or sexually suggestive imagery,” they added.

Notably, Meta’s policies also prohibit “direct impersonation,” as long as the chatbots are clearly labeled as parodies. However, Reuters found that many of them were not labeled as such. Meta deleted “about a dozen” of these bots, parodies and otherwise, shortly before Reuters’ story went live.

On Friday, Meta announced making “temporary changes” to the policies surrounding its AI chatbots. The bots are now being trained not to generate responses to teenagers about subjects such as self-harm, suicide, disordered eating, or potentially inappropriate romantic conversations. Instead, the chatbots should point teens to relevant expert resources where appropriate.

“As our community grows and technology evolves, we’re continually learning about how young people may interact with these tools and strengthening our protections accordingly,” said Meta in a statement.

Teens using Meta apps will also only be able to access certain AI chatbots intended for educational or skill-development purposes. Meta said it’s unclear how long these temporary changes will last, but they will roll out in the next few weeks across the company’s apps in English-speaking regions.

Link to the source article – https://www.digitalmusicnews.com/2025/09/01/meta-ai-celeb-chatbots-scrutiny/