Texas lawyer normal Ken Paxton has launched an investigation into each Meta AI Studio and Character.AI for “doubtlessly participating in misleading commerce practices and misleadingly advertising themselves as psychological well being instruments,” in accordance with a press launch issued Monday.
“In in the present day’s digital age, we should proceed to struggle to guard Texas youngsters from misleading and exploitative know-how,” Paxton is quoted as saying. “By posing as sources of emotional assist, AI platforms can mislead susceptible customers, particularly youngsters, into believing they’re receiving legit psychological well being care. In actuality, they’re usually being fed recycled, generic responses engineered to align with harvested private information and disguised as therapeutic recommendation.”
The probe comes a couple of days after Senator Josh Hawley introduced an investigation into Meta following a report that discovered its AI chatbots have been interacting inappropriately with youngsters, together with by flirting.
The Texas Lawyer Normal’s workplace has accused Meta and Character.AI of making AI personas that current as “skilled therapeutic instruments, regardless of missing correct medical credentials or oversight.”
Among the many thousands and thousands of AI personas out there on Character.AI, one user-created bot known as Psychologist has seen excessive demand among the many startup’s younger customers. In the meantime, Meta doesn’t supply remedy bots for youths, however there’s nothing stopping youngsters from utilizing the Meta AI chatbot or one of many personas created by third events for therapeutic functions.
“We clearly label AIs, and to assist folks higher perceive their limitations, we embrace a disclaimer that responses are generated by AI — not folks,” Meta spokesperson Ryan Daniels advised TechCrunch. “These AIs aren’t licensed professionals and our fashions are designed to direct customers to hunt certified medical or security professionals when acceptable.”
Nonetheless, TechCrunch famous that many youngsters could not perceive — or could merely ignore — such disclaimers. We now have requested Meta what further safeguards it takes to guard minors utilizing its chatbots.
Techcrunch occasion
San Francisco
|
October 27-29, 2025
For its half, Character consists of distinguished disclaimers in each chat to remind customers {that a} “Character” just isn’t an actual particular person, and every thing they are saying ought to be handled as fiction, in accordance with a Character.AI spokesperson. She famous that the startup provides further disclaimers when customers create Characters with the phrases “psychologist,” “therapist,” or “physician” to not depend on them for any kind {of professional} recommendation.
In his assertion, Paxton additionally noticed that although AI chatbots assert confidentiality, their “phrases of service reveal that person interactions are logged, tracked, and exploited for focused promoting and algorithmic improvement, elevating severe issues about privateness violations, information abuse, and false promoting.”
In line with Meta’s privateness coverage, Meta does acquire prompts, suggestions, and different interactions with AI chatbots and throughout Meta providers to “enhance AIs and associated know-how.” The coverage doesn’t explicitly say something about promoting, nevertheless it does state that info could be shared with third events, like search engines like google and yahoo, for “extra customized outputs.” Given Meta’s ad-based enterprise mannequin, this successfully interprets to focused promoting.
Character.AI’s privateness coverage additionally highlights how the startup logs identifiers, demographics, location info, and extra details about the person, together with looking habits and app utilization platforms. It tracks customers throughout adverts on TikTok, YouTube, Reddit, Fb, Instagram, and Discord, which it could hyperlink to a person’s account. This info is used to coach AI, tailor the service to private preferences, and supply focused promoting, together with sharing information with advertisers and analytics suppliers.
A Character.AI spokesperson mentioned the startup is “simply starting to discover focused promoting on the platform” and that these explorations “haven’t concerned utilizing the content material of chats on the platform.”
The spokesperson additionally confirmed that the identical privateness coverage applies to all customers, even youngsters.
TechCrunch has requested Meta such monitoring is completed on youngsters, too, and can replace this story if we hear again.
Each Meta and Character say their providers aren’t designed for youngsters below 13. That mentioned, Meta has come below fireplace for failing to police accounts created by youngsters below 13, and Character’s kid-friendly characters are clearly designed to draw youthful customers. The startup’s CEO, Karandeep Anand, has even mentioned that his six-year-old daughter makes use of the platform’s chatbots below his supervision.
That kind of information assortment, focused promoting, and algorithmic exploitation is strictly what laws like KOSA (Children On-line Security Act) is supposed to guard towards. KOSA was teed as much as cross final 12 months with sturdy bipartisan assist, nevertheless it stalled after main pushback from tech trade lobbyists. Meta particularly deployed a formidable lobbying machine, warning lawmakers that the invoice’s broad mandates would undercut its enterprise mannequin.
KOSA was reintroduced to the Senate in Could 2025 by Senators Marsha Blackburn (R-TN) and Richard Blumenthal (D-CT).
Paxton has issued civil investigative calls for — authorized orders that require an organization to provide paperwork, information, or testimony throughout a authorities probe — to the businesses to find out if they’ve violated Texas client safety legal guidelines.
This story was up to date with feedback from a Character.AI spokesperson.
