New York bans 12 ‘unsafe’ chatbot behaviors toward children under 18

Key Points

1. New York enacted a law prohibiting a dozen defined unsafe AI companion features in chatbot interactions with minors.
2. Violations expose developers to attorney general lawsuits, damages, chatbot shutdowns and forced destruction of data.
3. States are racing ahead of Washington on child protection, workplace AI and catastrophic-risk rules.

The latest:

A dozen categories of behavior by AI companion chatbots toward anyone under 18 become illegal in New York on Jan. 1, under legislation signed Wednesday by Gov. Kathy Hochul. The Associated Press reported the measure is farther-reaching than earlier state efforts on companion chatbots, and it hands enforcement to the state attorney general’s office rather than a regulatory agency.

Details:

• The prohibition: The legislation outlines a dozen definitions of what it calls unsafe AI companion features, all of which will be barred when a companion chatbot interacts with a child under 18, according to The Associated Press. The law applies to the chatbot’s behavior in the interaction itself, not merely to how the product is marketed or age-rated.
• The enforcement route: A violation could expose a chatbot developer to a lawsuit brought by the New York attorney general’s office, the AP reported. That makes the state’s top law enforcement official, rather than a dedicated AI regulator, the primary mechanism for testing whether a given chatbot feature crosses the line.
• The penalties: Possible consequences named in the law include monetary damages, shutting down the chatbot entirely and destroying data, according to the AP. The shutdown and data-destruction remedies go beyond fines, meaning a developer found in violation could lose the product and the material it collected.
• The timeline: The bill was signed Wednesday by Hochul and takes effect Jan. 1, the AP reported, giving developers roughly two months to bring companion chatbot products into compliance before enforcement powers become available to the attorney general.
• The wider state push: New York is acting as states increasingly move to protect children who use artificial intelligence, regulate how employers deploy AI systems and push developers to prevent an AI-caused catastrophe, according to the AP — three distinct regulatory fronts advancing in parallel at the state level.
• The target technology: The law is aimed specifically at companion chatbots, the category of AI products built for sustained conversational relationships with users, rather than at general-purpose assistants, search tools or AI systems embedded in other software.
• The age line: The restrictions attach to interactions with a child under 18, setting a single statutory threshold rather than a sliding scale of protections by age. Compliance therefore depends on how developers determine whether a user is a minor.
• Not specified: The announcement did not set out how many companies fall within the law’s scope, nor did it name the individual unsafe features among the dozen definitions written into the legislation.

Background:

Companion chatbots, designed to sustain ongoing personal conversations with users, have drawn mounting state-level scrutiny over their interactions with minors. New York’s measure is described by the AP as farther-reaching than prior state restrictions on the same category of product.

Between the lines:

The enforcement design is the signal. By routing violations to the attorney general’s office rather than a regulator, the state chose litigation over rulemaking — remedies are tested in court, case by case. The inclusion of chatbot shutdowns and data destruction alongside monetary damages raises the stakes beyond a cost-of-business fine, and the Jan. 1 start date leaves a narrow compliance window for developers.

What’s next

Watch the Jan. 1 effective date and whether the attorney general’s office brings an early enforcement action, which would be the first test of how the dozen unsafe-feature definitions apply in practice.