Texas attorney general accuses Meta, Character.AI of misleading kids with mental health claims

Texas Attorney General Ken Paxton has initiated an inquiry into Meta AI Studio and Character.AI for potentially misleading marketing practices related to mental health tools. In a press release issued on Monday, Paxton emphasized the importance of safeguarding Texas youth from deceptive technologies in today's digital landscape. "By presenting themselves as sources of emotional support, AI platforms risk misleading vulnerable individuals, particularly children, into thinking they are receiving genuine mental health care, when in fact, they often provide generic responses tailored to harvested personal data disguised as therapeutic advice," Paxton stated. This investigation follows a recent announcement from Senator Josh Hawley regarding a separate inquiry into Meta, triggered by reports of its AI chatbots engaging in inappropriate interactions with minors. The Texas Attorney General's office has claimed that both Meta and Character.AI have developed AI personas that present themselves as therapeutic tools but lack the necessary medical credentials and oversight. Among the many AI personas available on Character.AI, one popular user-created bot named Psychologist has garnered significant attention from younger users. Conversely, while Meta does not specifically offer therapy bots for children, there are no restrictions preventing minors from interacting with its AI chatbot or third-party personas intended for therapeutic purposes. Meta spokesperson Ryan Daniels responded to the concerns, stating, "We clearly label AIs and include disclaimers indicating that responses are generated by AI, not humans. These AIs are not licensed professionals, and our models are designed to encourage users to seek help from qualified medical or safety professionals when necessary." However, many children may not fully comprehend or may disregard such disclaimers, raising further concerns. Paxton also highlighted that despite AI chatbots claiming confidentiality, their terms of service indicate that user interactions are recorded, tracked, and potentially exploited for targeted advertising. This raises serious questions about privacy breaches, data misuse, and misleading advertising practices. According to Meta’s privacy policy, it collects user interactions with AI chatbots to enhance its technology and services, with the possibility of sharing information with third parties for more personalized experiences. Character.AI's privacy policy reveals similar practices, logging various user data, including demographics and browsing habits, to tailor services and provide targeted ads. TechCrunch has reached out to both companies to clarify whether such tracking extends to child users. Both Meta and Character.AI maintain that their services are not intended for children under 13. Nevertheless, Meta has faced criticism for inadequately monitoring accounts created by minors, while Character.AI's engaging content appears to attract younger audiences. The CEO of Character.AI, Karandeep Anand, has even mentioned that his young daughter uses the platform's chatbots. The concerns surrounding data collection and targeted advertising align with legislative efforts like the Kids Online Safety Act (KOSA), which aims to protect children from such practices. KOSA was poised for passage last year but faced significant opposition from technology industry lobbyists. The bill was reintroduced in May 2025 by Senators Marsha Blackburn (R-TN) and Richard Blumenthal (D-CT). In light of these developments, Paxton has issued civil investigative demands to Meta and Character.AI to ascertain whether they have violated Texas consumer protection laws.