Iowa law puts labeling, AI reminder requirements on conversational chatbots

New legislation in Iowa puts guidelines on AI chatbots. Senate File 2417 passed unanimously in the House and Senate, meaning chatbots like Claude and ChatGPT must disclose to young users that they are not human, and refrain from encouraging kids to extend conversations.

Moreover, a labeling requirement says chatbots must clearly label AI content and distinguish between things created by humans and generative models.

As quoted in the Des Moines Register, Republican state Senator Kara Warme says “the main goal of SF 2417 is to make chatbots safer for users, especially children, by making them more transparent and less deceptive.”

“We see a lot of value in human connection, and so we want to make sure it’s clear whether you are interacting with a human or something that’s technology or computer based.”

The regulation is not extraordinary – many jurisdictions are considering restrictions on chatbots, as pressure mounts from parents and civil society for regulation to address the problem. But the sentiment in Warme’s statement is only now beginning to surface as a political issue – namely, that human connection is something that must be preserved and cannot be replaced by AI.

No mental health counseling, no sex from chatbots

The more extreme risks of going too deep with chatbots have been well documented, and Iowa’s law prohibits them from offering mental or psychological health counseling or advice, and requires them to refer any user presenting risks of suicide or self harm to professionals. Chatbots can’t produce sexually explicit visual media, encourage children to engage in sexually explicit acts, or sexually objectify minors.

The section of the law addressing the not-human reminder feature specifies that “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.”

Are we approaching legislated outdoor time?

The law went into effect July 1, but the state won’t begin enforcing it until July 1, 2027, to allow operators a transitional window. Lawmakers have been careful to emphasize that the law applies only to conversational chatbots whose “main purpose is human interaction,” and not broader applications of machine learning. Violations come with civil penalties of up to $500,000 per operator.

The legislation was prompted in part by several cases involving Iowans who used AI to make child pornography. Warme has encouraged parents to talk with their kids and their friends about the risks inherent in AI chatbots. “And to take breaks from technology that they’ve been using, to make sure that they have real life interactions and a chance to get outside as well.”

We have reached the point in our modern technological journey at which even the politicians are telling us to touch grass. Chatbots, while extremely popular, are not yet as established as social media platforms, and could end up being an innovation that, ultimately, humanity decides is no substitute for the real thing.

Article Topics

age verification  |  chatbots  |  children  |  Iowa  |  legislation  |  United States