A 10379: Prohibits artificial intelligence companions from using features which are considered unsafe for minors - New York

A10379B → A10379C · 4 edits

MODERATE

The bill was amended to broaden its scope from regulating only 'chatbots' to covering a wider range of 'artificial intelligence companions.' This change updates the terminology throughout the text to align with a more comprehensive definition of AI systems that interact with users, ensuring the law applies to various forms of AI companionship rather than just traditional chat interfaces.

Scope change

The bill's scope expanded from specifically targeting 'advanced chatbots' to covering 'AI companions,' which includes a broader class of generative AI systems designed to provide ongoing, adaptive responses to users.

DEFINITION

Replaced the term 'advanced chatbot' with 'AI companion' and 'Covered AI companion' to encompass a wider variety of AI systems beyond simple text-based chat tools.

Added new definitions for 'Artificial intelligence,' 'AI model,' and 'Generative artificial intelligence' to clarify the technology covered by the law.

REQUIREMENT

Updated all references to prohibited features (such as simulating human relationships or emotional appeals) to apply to 'AI companions' instead of 'chatbots,' ensuring the safety restrictions apply to the broader category of AI systems.

ELIGIBILITY

Changed the term 'chatbot operator' to 'Operator' to reflect that the regulations apply to any entity providing these AI companions, not just those operating chatbots.

A10379A → A10379B · 3 edits

MINOR

The bill was amended to reorganize and clarify the definition of 'unsafe chatbot features' by merging several sub-points into broader categories. A new definition for 'deceptive as to minors' was added to specifically address misleading outputs aimed at children. The bill was also re-introduced with additional sponsors and a new procedural status indicating it has been amended and re-committed.

DEFINITION

The definition of 'unsafe chatbot features' was consolidated and renumbered. Sub-points regarding companionship simulation were merged into a single clause about implying a personal relationship or authority figure role, and a new clause was added to prohibit outputs that are deceptive to minors about the chatbot's non-sentient nature.

A new definition for 'deceptive as to minors' was introduced to explicitly cover outputs that create an atmosphere conducive to fraud or deceive credulous minors.

PROCEDURAL

The bill's procedural history was updated to reflect additional sponsors and a new status where the committee has reported the bill back with amendments.

A10379 → A10379A · 4 edits

MODERATE

The bill was amended to add four new sponsors and update the procedural history to reflect that the committee has discharged the bill, amended it, and ordered it reprinted. The substantive text of the legislation, including the definitions of unsafe AI features and the prohibitions on using them with minors, remains unchanged.

Scope change

The scope of the bill's applicability remains the same; no changes were made to the laws governing AI chatbots or their restrictions.

ELIGIBILITY

Added four new sponsors to the bill: Shimsky, Lee, McMahon, Stirpe, Conrad, Hevesi, Lunsford, Burdick, and Burroughs.

Updated the procedural history to indicate the bill was discharged from committee, amended, and ordered reprinted before being recommitted.

TECHNICAL

Adjusted the bill number format from '10379' to '10379--A' to reflect the amended version.

Minor formatting adjustments were made to the spacing and line breaks in the explanatory notes and section headers.

Prohibits artificial intelligence companions from using features which are considered unsafe for minors

What changed between versions

How they voted

Actions timeline

Sponsors