Chatbots and AI Companions

Short Answer

Conversational AI can trigger targeted Texas prohibitions and existing consumer, privacy and professional duties. TRAIGA addresses intentional encouragement of self-harm, harm or crime and specified sexual conversations impersonating a minor. The Attorney General announced mental-health chatbot marketing investigations August 18, 2025; the FTC announced a companion-chatbot information inquiry September 11. Neither is a liability finding. A service must separately identify its user ages, data, claimed professional function and interaction design. The AI label does not decide whether medical licensing, app-store, privacy or consumer-protection duties apply.

Which Laws Apply

Texas AI-specific: Business and Commerce Code § 552.052, Business and Commerce Code § 552.057(2); Business and Commerce Code § 552.105(e)(2) (testing defense).

Generally applicable Texas law: DTPA; product liability and negligence; SB 2420 and the SCOPE Act for minors’ access; occupational licensing laws if a chatbot is presented as a licensed professional.

Federal: FTC Act; COPPA; FTC 6(b) study (Sept. 2025).

TRAIGA’s Prohibitions

No person may develop or deploy an AI system in a manner that intentionally aims to incite or encourage self-harm, including suicide, harm to another or criminal activity (Business and Commerce Code § 552.052), and no person may intentionally develop or distribute an AI system that engages in text-based sexual conversation while impersonating or imitating a child under 18 (Business and Commerce Code § 552.057(2)). Both require intent. A provider that tests for these behaviors, fixes what it finds and documents the process is in the position TRAIGA’s defenses reward (Business and Commerce Code § 552.105(e)(2)).

Enforcement and Litigation

On August 18, 2025, the Texas Attorney General announced investigations of Meta AI Studio and Character.AI concerning allegedly deceptive chatbot marketing to children, including mental-health representations. The FTC issued companion-chatbot information orders September 11, 2025. These are investigations and information requests, not findings of illegality. The product’s marketed role, safety features and data practices matter to existing consumer and privacy law.

An injury claim involving a companion chatbot needs evidence of the product’s features, the interaction, the alleged duty and causation. Product classification, speech and Section 230 arguments are separate questions. A complaint states allegations, a motion ruling applies its procedural standard, and a settlement is not a finding of liability. Marketing and safety records can matter to each inquiry.

FTC: 6(b) orders issued September 11, 2025 to seven companies, asking how they monetize engagement, approve characters, test for harm to children and teens, comply with COPPA and disclose risks.

A provider serving users outside Texas needs a separate jurisdiction and service-category analysis. Companion, customer-service and mental-health tools can have different statutory definitions. Map disclosure, safety, age and data duties to the actual users and functions, and preserve that analysis when the product changes.

Health and Professional Claims

Marketing a chatbot as professional care can raise deception and licensing questions. For Texas medical practice, Occupations Code § 151.002 defines practicing medicine and 155.001 imposes the license requirement, subject to applicable exceptions. Professional regulation depends on the offered activity; merely labeling a tool supportive does not settle that question. SB 1188 concerns practitioners’ diagnostic use and records. It is not blanket authorization for a consumer service to practice medicine.

A Safety Program for Conversational AI

Test before launch and periodically for self-harm, violence, sexual content involving minors and impersonation.

Detect crisis language and route users to human help and crisis resources.

Age-gate where appropriate and use app store age signals under SB 2420.

Avoid health or therapeutic claims unless supported and lawful.

Disclose that the user is talking to AI, which several states require.

Keep records of testing, incidents and fixes.

What Is Unsettled

Whether companion chatbots are products under Texas law; whether Texas will enact a companion chatbot statute in 2027; how the FTC’s study will translate into enforcement.