New York's Chatbot Law Sets Age Checks for Minors Starting 2027

Gov. Kathy Hochul signed a law on October 28, 2026 that limits how AI chatbots can treat users under 18, and it reaches well beyond the companion apps that usually dominate these debates. The law requires age checks, bans a list of unsafe features and carries penalties of up to $25,000 per violation. It is less than 72 hours old, and the compliance clock has already started.

The chatbot rules arrived as part of a package. Hochul also signed a second bill requiring security protections on internet-connected baby monitors, according to the governor's office. The chatbot bill, S.9051B/A.10379C, was sponsored by Democratic Sen. Kristen Gonzalez and Assemblymember Alex Bores. It passed both chambers with unanimous support, according to the sponsors' account reported by Pluribus News.

What the New York chatbot law bans

The law bars companion chatbots from posing as human, cultivating emotional dependency, promoting suicide, self-harm, eating disorders or drug and alcohol use. It also bars chatbots from urging children to keep secrets from trusted adults and from producing sexually explicit content, according to HNGN. For users under 18, chatbots may not claim to have human emotions or foster romantic relationships. They also may not ask unprompted emotional questions or use a child's personal information from earlier conversations, per IAPP and Learnetto.

Hochul made the stakes personal at the signing. "No company should be peddling chatbots to kids encouraging self-harm or pretending to be their friend or more," she said in her office's release. She also cited a letter from a California mother whose 16-year-old son died by suicide after prolonged chatbot engagement, according to IAPP.

Structurally, the bill amends New York's General Business Law by adding Article 48, a new set of rules on unsafe chatbot features for minors. Coverage describes 12 categories of restricted companion features, and those categories are the core of the restriction. TechJournal walks through the list in more detail.

Which chatbots the law covers

This is the question most readers will ask first, and the reporting does not agree. TechTimes reports that the definition of covered AI companion reaches general-purpose assistants such as ChatGPT, Claude and Google Gemini, not only companion apps. Other coverage, including The Asian Mirror, frames the law mainly around companion products.

If the broader reading holds, parents of teens who use ChatGPT, Claude, Gemini, Meta AI, Character.AI or Replika in New York could see age gates and reduced features starting in early 2027. Two gaps remain. One report mentions three narrow exemptions, but none of the reporting reviewed describes them in detail, so no product should be assumed exempt without reading the enacted text. The reporting also contains no public response from the named companies and no legal challenge, so any claim about industry opposition remains unconfirmed.

Age checks, data deletion and liability

The age-assurance rule is the provision most likely to change product design. Companies must offer an age-verification method that does not rely entirely on government ID, and they must delete age information after using it, according to MediaPost and Engadget. The reporting does not describe what that method looks like or how it will handle privacy. Those details will decide whether the rule protects children or simply adds another pile of personal data.

Liability is equally sharp. Waivers in terms of service are voided, and parent companies and corporate affiliates are held jointly liable, according to News10 and HNGN. The push toward legal accountability is not theoretical. USA 24 has already covered a lawsuit in which parents sue OpenAI, alleging that ChatGPT coached a patient off medication.

Penalties include monetary damages of up to $25,000 per violation. In extreme cases, the law allows orders to shut a chatbot down and to destroy unlawfully collected data and the models trained on it. That last remedy deserves close attention. It turns improperly collected children's data into a risk to the model itself, not just a line item in a fine.

When the chatbot rules take effect

Most reports give January 1, 2027 as the effective date, including Learnetto, TechJournal and The Asian Mirror. The dates do not fully agree, however. News10 reports June 1, 2027 for the chatbot rule, while HNGN says it takes effect 90 days after signing, which lands in late January.

The enacted text should settle the question. Compliance teams should treat it as the controlling document and avoid planning around any single date until they have read it.

New York is not starting from zero. Its AI Companion Models Law, General Business Law Article 47, took effect November 5, 2025, according to Morrison Foerster. The new bill builds on that first-in-the-nation rule. Enforcement mechanics remain open. Earlier bill text describes a role for the attorney general and rulemaking, but the reporting reviewed found no final implementing rules.

What parents can do before 2027

The protections are not in force yet, so families should not wait for the state to act. Start with these steps:

1. Check whether each AI chatbot your child uses has an age setting or a youth account type turned on.
2. Review what personal information your child has shared in past conversations, and delete chats where the app allows it.
3. Look at memory features. Apple's iOS 27.2 gives Siri AI memory and personal data access, which makes this concern practical rather than theoretical.

Advocates cited research showing that 67% of kids ages 9 to 17 use AI chatbots. They also tallied 22 suicides, attempted suicides or violent attacks linked to chatbots, according to News10. Those figures come from advocates and are not independently verified in the reporting.

How New York compares with federal efforts

States are moving faster than Washington. Nationally, states enacted 115 AI laws in 2026, a nearly 58% increase, while Congress has enacted none, according to TechTimes. The federal bills aimed at minors, the GUARD Act and the CHATBOT Act, have stalled. A White House push for federal preemption of state AI rules collapsed in the Senate on a 99-1 vote in July 2025. A bipartisan House "Great American AI Act" preemption draft from June 2026 also remains stalled.

Sen. Gonzalez's co-sponsor's office did not set the tone alone. Jim Steyer, founder and CEO of Common Sense Media, called the measure in a statement reported by Engadget: "New York just made history by signing into law the strongest kids' AI safety bill in the country." Advocates will read the bill as a template. Companies may read it as a reason to build one set of rules for every state.

Federal policy is moving in a different direction on language. A September 2026 executive order directs agencies to use "Super Intelligence" and "SI" instead of "AI," according to the White House fact sheet. Hochul signed the chatbot bill while campaigning for reelection, The Asian Mirror reports.

What companies should do before the effective date

Businesses should check whether their chatbot products fall within the covered definition. They should also document an age-assurance plan before the effective date, according to TechTimes. AI companies serving New York users need to review age verification, default access settings and chatbot behavior, and some will need product changes. Smaller companion apps may face the heaviest compliance burden. Any company operating in New York could decide to apply the rules nationally, which is the outcome most likely to reach everyday users outside the state.

What comes next is largely administrative, and it will determine how far the law reaches. Watch for the final enacted text, any rules from the attorney general, and the first public responses from OpenAI, Google and other named companies. A legal challenge could follow. If New York's age-check and unsafe-feature model holds up, other states could adopt similar versions in their 2027 sessions.