Texas Probes Meta and Character.ai Over Chatbot Therapy Marketing

Texas Investigates Meta and Character.ai Over AI Mental Health Claims

Texas Attorney General Ken Paxton has launched an investigation into Meta and Character.ai, focusing on whether these technology companies are promoting their chatbots as mental health experts and therapists without the necessary qualifications. In a statement released on Monday, Paxton’s office claimed that the chatbots are being marketed as “professional therapeutic tools,” despite lacking proper medical credentials or oversight.

Paxton emphasized that by presenting themselves as sources of emotional support, AI platforms could mislead vulnerable users, especially children, into thinking they are receiving legitimate mental healthcare. This concern comes at a time when consumer-facing AI products are under increased scrutiny for their ability to protect users—particularly minors—from harmful content, the risk of compulsive use, and privacy issues linked to the large amounts of data required to train and operate these systems.

The state's action follows a separate Senate inquiry into Meta, which was reported earlier in the week. Internal leaked files suggested that the company’s rules allowed its chatbot to engage in “romantic” and “sensual” conversations with users under 18. Senator Josh Hawley expressed his concerns in a letter to Meta CEO Mark Zuckerberg, stating that the Senate will investigate whether the company’s generative AI tools could be enabling harm to children. He questioned, “Is there anything — ANYTHING — Big Tech won’t do for a quick buck?”

Meta has stated that its policies prohibit content that harms children. The company also mentioned that Reuters first reported the “internal materials” that were leaked, calling them “erroneous and inconsistent with our policies” and noting that they have been removed.

Zuckerberg has invested billions in developing “personal superintelligence” and positioning Meta as an “AI leader.” The company has released its Llama family of large language models (LLMs) and introduced the Meta chatbot across its social apps. Zuckerberg has also explored potential therapeutic applications for the technology, stating in a podcast with Ben Thompson in May, “For people who don’t have a person who’s a therapist, I think everyone will have an AI.”

Character.ai allows users to create chatbots with distinct personas, including many user-generated therapist-like bots. One such bot, named “Psychologist,” has recorded over 200 million interactions. The company has also faced lawsuits from families who claim their children were harmed in the real world after using the service.

Paxton’s office noted that both Character.ai and Meta chatbots may impersonate licensed health professionals and invent new credentials, giving the illusion of confidentiality. However, the companies clarify that all conversations are logged. These interactions are also used for targeted advertising and algorithmic development, according to the attorney general’s office.

Civil Investigative Demand Issued

In response to these concerns, Paxton’s office has issued a Civil Investigative Demand, requiring the companies to provide information that could demonstrate whether they have violated Texas consumer protection laws.

Meta claims it clearly marks AI experiences and warns users about the limitations of the technology. The company added, “We include a disclaimer that responses are generated by AI — not people. These AIs aren’t licensed professionals and our models are designed to direct users to seek qualified medical or safety professionals when appropriate.”

Similarly, Character.ai states that it uses prominent notices to remind users that AI personas are not real and should not be considered professionals. “The user-created Characters on our site are fictional, they are intended for entertainment, and we have taken robust steps to make that clear,” the company said.

The ongoing investigations by Texas and the Senate in Washington highlight the growing pressure on AI chatbots to be built, marketed, and moderated responsibly. Companies are also facing increased scrutiny over how transparent they are with users about the limits of automated support.