House File 2507 -H-8357 IntroducedAmend HOUSEHouse FILEFile 2507 BYas COMMITTEEfollows: ON HEALTH AND HUMAN SERVICES (SUCCESSOR TO HSB 611) A BILL FOR An Act establishing requirements and guidelines for conversational AI services, and providing civil penalties, and including applicability provisions.

BE1. IT ENACTED BY THE GENERAL ASSEMBLY OF THE STATE OF IOWA:

TLSBBy 5347HVstriking (3)everything 91after dg/jhthe H.F.enacting clause and inserting:

2507<Section Section 1.

“Operator” HF 2507.3944 (1) 91 (amending this HF 2507 to CONFORM to SF 2417) dg/jh 1/4 -1- does not include a mobile device application store or a search engine solely because the mobile device application store or a search engine provides access to a conversational AI service.

LSB5. 5347HV (3) 91 -1- dg/jh 1/6 H.F.

2507 5.

An operator shall institute reasonable measures to prevent the operator’s conversational AI service from HF 2507.3944 (1) 91 (amending this HF 2507 to CONFORM to SF 2417) dg/jh 2/4 -2- generating statements that would lead a reasonable individual to believe that the individual is interacting with a human, including but not limited to all of the following:

LSBa. 5347HV (3) 91 -2- dg/jh 2/6 H.F.

2507 a.

An operator shall offer tools for the parent or guardian of a minor account holder to manage the minor account holder’s privacy and account settings ifas theappropriate minorbased hason additionalrelevant riskrisks. factors identified by the attorney general by rule.

An operator shall clearly and conspicuously disclose using a persistent visible disclaimer, or a disclaimer that appears after every three hours of continuous interaction with the operator’s conversational AI service, that the operator’s conversational AI service is artificial intelligence if a reasonable individual interacting with the conversational AI service would believe that the individual is interacting with a human.

An operator shall adopt protocols for the operator’s conversational AI service for responding to user prompts regarding suicidal ideation or self-harm that includes but is not limited to making reasonable efforts to refer the user to HF 2507.3944 (1) 91 (amending this HF 2507 to CONFORM to SF 2417) dg/jh 3/4 -3- crisis service providers such as a suicide hotline, crisis text line, or other appropriate crisis service.

An operator shall not knowingly and intentionally cause or program a conversational AI service to make a representation or LSBstatement 5347HVthat (3)would 91lead -3-a dg/jhreasonable 3/6individual H.F.to believe that the conversational AI service is designed to provide professional psychology or behavioral health services that an individual would require licensure under chapter 154B or 154D to provide.

2507 statement that would lead a reasonable individual to believe that the conversational AI service is designed to provide professional psychology or behavioral health services that an individual would require licensure under chapter 154B or 154D to provide.

ThisA chaptercivil shallpenalty notcollected beunder construedthis tosection createshall abe privatedeposited rightinto ofthe actiongeneral underfund thisof chapterthe orstate. any other law.

This chapter shall not be construed to create a private right of action under this chapter or any other law.

5.

This Act applies July 1, 2027.2027.> ______________________________ HARRIS of Appanoose HF 2507.3944 (1) 91 (amending this HF 2507 to CONFORM to SF 2417) dg/jh 4/4 -4-

EXPLANATION The inclusion of this explanation does not constitute agreement with the explanation’s substance by the members of the general assembly.

This bill relates to requirements and guidelines for conversational AI services (AI services).

The bill defines “conversational AI service” as an artificial intelligence, available by software application, web interface, or computer program, that is accessible to the general public and that has the primary purpose of simulating human conversation and interaction through text, audio communication, or visual communication.

The bill lists exceptions to the definition of AI service.

The bill defines “operator” as a person who develops and makes an AI service available to the public.

The bill lists LSB 5347HV (3) 91 -4- dg/jh 4/6 H.F.

2507 exceptions to the definition of “operator”.

The bill also defines “account holder”, “minor”, “sexually explicit conduct”, and “visual depiction”.

The bill requires an operator to clearly and conspicuously disclose to a minor account holder (minor) that the minor is interacting with artificial intelligence through a persistent visible disclaimer or a disclaimer that appears at the beginning of each interaction between the operator’s AI service and a minor and at least once every three hours of continuous interaction between the minor and the AI service.

The bill prohibits an operator from providing a minor with points or similar rewards at unpredictable intervals with the intent to encourage increased engagement with the operator’s AI service.

The bill requires an operator to prevent the operator’s AI service from taking certain actions as described in the bill for a minor.

The bill requires an operator to institute reasonable measures to prevent the operator’s AI service from leading a reasonable individual to believe the individual is interacting with a human.

If the AI service would lead a reasonable individual to believe the individual is interacting with a human, the operator must clearly and conspicuously disclose using a persistent visible disclaimer that the AI service is artificial intelligence.

The bill requires, in circumstances detailed in the bill, an operator to provide tools to a minor, a minor’s parent or guardian, or both that will allow the individual to manage the minor’s privacy and account settings.

The bill requires an operator to adopt protocols for the operator’s AI service that would respond as detailed in the bill to user prompts of suicidal ideation or self-harm.

The bill prohibits an operator from knowingly and intentionally causing or programming an AI service to make a representation or statement that would lead a reasonable person LSB 5347HV (3) 91 -5- dg/jh 5/6 H.F.

2507 to believe the AI service is designed to provide professional psychology services or behavioral health services that would require licensure.

The bill requires the attorney general to enforce the bill and adopt rules to administer the bill.

An operator in violation of the bill is subject to an injunction and the greater of actual damages or a civil penalty of $1,000 per violation, up to a maximum of $500,000 per operator.

The bill does not create a private right of action and shall not be construed to make a developer of an artificial intelligence model liable solely because a third party used the developer’s artificial intelligence model to create or train an AI service.

The bill applies July 1, 2027.