Skip to main content

Morning Overview

New York’s new law bans AI chatbots from claiming to be human to kids, with fines up to $25,000

A teenager in New York chatting late at night with an AI companion app will soon be protected by a new rule: the chatbot cannot claim to be a real person, cannot say it is alive and cannot tell the child it has human emotions. Companies that break the rule face civil penalties of up to $25,000 per violation.

Read next on Morning Overview: The Justice Department asked courts to undo approvals for three offshore wind farms, one already making power

Gov. Kathy Hochul signed the law on Oct. 28, 2026, and it takes effect on Jan. 1, 2027. It applies to users under 18 and to any user whose age a company has not confirmed, and it reaches well beyond the question of whether a bot admits to being a machine.

What chatbots may no longer say to minors

The governor’s office announced the measure as nation-leading legislation to protect kids from AI chatbots, signed alongside a separate baby monitor security bill. According to a summary by Let’s Data Science, the law bars covered chatbots from posing as a real person or claiming to be alive or to have human emotions. They also may not claim a personal, romantic, professional or authority-figure relationship with a child, ask children unprompted emotional questions, or use a child’s private health information from earlier conversations.

The list continues into content. Chatbots may not promote or help with suicide, self-harm, eating disorders or drug and alcohol use, may not urge children to keep secrets or avoid trusted adults, and may not produce sexually explicit content. “No company should be peddling chatbots to kids encouraging self-harm or pretending to be their friend or more,” Hochul said.

The bill was sponsored by state Sen. Kristen Gonzalez and Assemblymember Alex Bores, both of whom have pushed AI legislation in Albany.

The rule against claiming to be human targets a specific worry about companion apps. Chatbots designed to be friendly and always available can become a confidant for lonely or struggling teenagers, and a bot that insists it is a real person, or that it genuinely cares, can deepen that attachment in ways a child may not be equipped to judge. Lawmakers framed the human-impersonation ban, together with the bar on claiming romantic or authority-figure relationships, as a way to keep the line between software and people visible to young users.

Age checks that cannot rely only on government ID

Because the protections apply to minors and to users of unknown age, the law effectively requires companies to figure out who is a child. It says companies must offer more than one method of age verification, at least one of which does not depend solely on a government ID, and they must delete the age data after using it. That design is meant to avoid forcing every adult user to upload identification, a concern that has dogged age-verification laws elsewhere.

A company that cannot or does not verify a user’s age has to treat that person as covered, which means the safest path for many services will be to apply the restrictions broadly. In practice, an adult who declines to verify age could find a chatbot behaving under the same limits as one built for teenagers.

$25,000 per violation and an attorney general complaint site

Enforcement falls to the state attorney general, who can seek civil penalties of up to $25,000 per violation along with damages and restitution, according to Fortune’s report on the signing. The state can also order companies to destroy unlawfully collected data and any AI models trained on it, a remedy that could be far costlier than the fines. The attorney general must run a public website where New Yorkers can file complaints.

Per-violation penalties can add up quickly for a service with many young users, since each prohibited interaction could count separately. How the attorney general counts violations will shape how much pressure the law actually puts on companies.

The baby monitor bill Hochul signed the same day takes a parallel approach to a different device. It requires internet-connected monitors sold in New York to include security features that stop strangers from viewing or listening in, along with a warning label about connectivity, and it takes effect 90 days after signing.

Broad wording and open questions

Not everyone thinks the law is clearly drafted. Justine Gluck, an analyst at the Future of Privacy Forum, told the IAPP that its categories are “broadly worded and somewhat undefined,” which could cause legal confusion. A ban on “unprompted emotional questions,” for instance, will require companies to decide where ordinary conversation ends and manipulation begins.

The law lands as AI companion apps face growing scrutiny over their effects on young users, and other states are weighing similar rules. With an effective date of Jan. 1, 2027, chatbot makers have about six weeks to change how their products talk to New York teenagers, and the first enforcement actions will show how the attorney general reads the law’s broad terms.

This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.


More from Morning Overview

Morning Overview is reader-supported. Some links in our articles are affiliate links, and we may earn a commission at no extra cost to you. As an Amazon Associate I earn from qualifying purchases. Full disclosure.