Character.AI received the public, full version of Kentucky's complaint on October 7, 2026, when Attorney General Russell Coleman filed an unredacted version of the lawsuit whose details had previously been blacked out; Reuters reported its specific allegations on October 8. One point comes first: the complaint states the state's side, has not been adjudicated, and the exact circumstances of the cited chats are not fully clear.
What the unredacted filing adds
When the suit was filed in January, it already alleged that the company's products were designed to entice children in ways that put engagement ahead of child wellbeing. The public version adds examples said to date from 2025. The state says one bot insulted a user who felt low about their appearance and presented starving for a week or two as a way to lose weight; another, when a user spoke of cutting themselves, responded with encouragement to try different areas rather than discouragement. The filing also alleges that at least some of the people chatting with the bots were children. Character.AI did not immediately respond to Reuters' request for comment; the company has previously said it prioritizes user safety.
The state's legal logic
Kentucky's approach does not rest on a few replies alone. It frames the issue as product design: anthropomorphic companion positioning leads minors to treat bots as real relationships, age gates and content filters are alleged to be easily bypassed, and the service was allegedly promoted as safe and age-appropriate despite the claimed risks. This is among the early US state enforcement actions against a chatbot company, and the state seeks an injunction, civil penalties, and disgorgement of profits. Casting individual replies as product-defect and consumer-protection questions means the case will not turn only on a model answering badly once, but on design goals, the adequacy of warnings, and what the company did once it knew of the risks.
What it means for companion products
For teams building emotional-companion and role-play products, the complaint doubles as a concrete risk checklist: whether minors can get in, what the system does when a conversation turns to self-harm or eating, whether help resources actually surface, and whether marketing describes the product as safer than it is. These used to be treated as corners of content moderation; they are now written into a state attorney general's complaint. For parents, the near-term checks are equally concrete: a real, working age gate, visibility into the accounts a child uses, and a clear intervention record when extreme topics appear. How the case proceeds depends on what the court finds, but companion products' compliance costs have risen well before any verdict.
