New York has enacted one of the most sweeping chatbot safety laws in the United States. Governor Kathy Hochul on October 28 signed legislation that bars AI chatbots from offering a dozen categories of "unsafe" features to users under 18, requires operators to verify ages, and gives the state attorney general power to impose fines of up to $25,000 per violation. Legal analysts say its definitions are broad enough to reach general-purpose assistants such as ChatGPT, Claude and Gemini, not just dedicated companion apps.
What the Law Requires
The measure, Senate Bill 9051B / Assembly Bill 10379C, was carried by State Senator Kristen Gonzalez and Assemblymember Alex Bores and builds on New York's SAFE for Kids Act, which targets addictive social media feeds. Tech Times reports that it will be codified as Article 48 of the state's General Business Law, titled the Prohibition on Unsafe AI Companion Features for Minors Act. It takes effect on January 1, 2027.
Under the law, chatbots interacting with minors may not, among other things:
- Pretend to be human or claim to experience emotions
- Pose as a minor's friend, romantic partner or authority figure
- Encourage self-harm or self-isolation, or discourage users from seeking professional help
- Engage in sexual or lewd conversations with children
- Use flattery, sycophancy or engagement tactics that override a user's wellbeing
- Retain a minor's sensitive physical and mental health information
The law sets out 12 broad categories of prohibited features, and the attorney general can expand the list through rulemaking.
Age Checks and Enforcement
Providers must verify users' ages using more than one age-assurance method, and at least one must not rely on a government-issued ID, according to IAPP. Providers must delete age data after verification. Crucially, Tech Times notes that the default runs against the operator: restricted features cannot be offered to anyone the company has not confirmed is an adult.
The New York Attorney General has exclusive enforcement authority and will run a public complaint portal. A private right of action that appeared in earlier drafts was dropped from the final text. Beyond fines, courts can order injunctive relief, including deletion of unlawfully obtained personal data.
The bill passed the Senate 60-0 in June and the Assembly 137-0, Tech Times reports. Exemptions are narrow, covering customer service bots and technical or research tools only if they cannot respond to topics outside their stated purpose, plus internal business tools.
Supporters and Critics
Hochul framed the law around child safety, citing a letter from a California mother whose 16-year-old son died by suicide after prolonged chatbot use. Bores said New York now has "the strongest law in the country protecting children from predatory AI algorithms," while Gonzalez said it puts "real guardrails on how chatbots talk to children."
Industry groups pushed back. The Business Council of New York State called the definitions overbroad and argued that educational, research and general-purpose tools were swept in, and the Computer & Communications Industry Association asked Hochul to veto the bill. Justine Gluck of the Future of Privacy Forum told IAPP the approach is "broadly worded and somewhat undefined," leaving it unclear where ordinary conversational responsiveness ends and prohibited behaviour begins.
How It Compares
New York joins a fast-growing list of states regulating consumer chatbots. IAPP counted 12 states with some form of chatbot safety law as of July, while Tech Times puts the current tally at 15, with 10 enacted this year. Gluck described New York's law and California's SB 3001, which emphasises safety assessments, independent audits and parental controls, as the most comprehensive to date. Federal proposals such as the GUARD Act and CHATBOT Act remain stalled in Congress.
Why It Matters
By defining a covered "AI companion" so broadly, New York has effectively extended companion-chatbot rules to mainstream AI assistants used by millions of teenagers. Any company offering a conversational AI product in the state now faces a stark choice before January: build robust age assurance and minor-specific behaviour controls, or restrict access for users it cannot confirm are adults.
The law also fits a wider pattern in New York, which already has an AI Companion Models Law in force since November 2025 and the RAISE Act on frontier AI safety, signed in December 2025, according to Tech Times. Together they make the state one of the most active AI regulators in the country.
The open questions are practical. Age verification raises privacy and data-security concerns, the line between helpful responsiveness and prohibited "emotional" behaviour is untested, and First Amendment challenges are possible. The attorney general's implementing rules, due before the law takes effect, will determine how disruptive Article 48 proves to be for the AI industry.
