A bill headed to the governor’s desk sets up new guardrails for chatbots that interact with minors in Iowa.
A chatbot must be programmed to disclose it’s not a human at the start of any online chat with an Iowa minor and it must remind the user it’s not human every three hours. The first chatbot was created in the 1960s, but Representative Austin Harris of Moulton said the recent surge of chatbots fueled by artificial intelligence have created some troubling situations. “Just Google it,” Harris said. “There have been several instances where AI chatbots that have been seeking mental health advice and encouraging them to commit self-harm.”
Representative Aime Wichtendahl of Hiawatha said the bill provides common sense protections. “We, as a state, are only starting are only beginning to understand the negative impacts that generative AI can have on our children,” Wichtendahl said.
Representative Beth Wessel-Kroeschell of Ames said while federal regulation is needed because these technologies don’t stop at state borders, Iowa shouldn’t wait for congress to act. “Conversational AI is evolving rapidly and the risks are real,” Wessel-Kroeschell said.
Senator Kara Warme of Ames indicated at least five other states are considering similar laws. “It’s an emerging need. Technology is valuable, but so is human connection,” Warme said. “And this is timely regulation to protect Iowans.”
The bill requires a chatbot to refer an Iowa child to a suicide hotline where they can speak with a human if the child brings up mental health issues. The bill creates a $1000 fine for a single violation of the law. If there are multiple violations, the maximum fine could be half a million dollars.