On Thursday, Reps. George Whitesides (D-CA), Mike Kennedy (R-UT), Doris Matsui (D-CA) and Mariannette Miller-Meeks (R-IA) introduced the bipartisan Protecting Kids from Human-Like Chatbots Act, legislation requiring developers to design chatbots to support a non-personified setting and make it the default for any user they reasonably know is a minor. Building off of ARI’s non-personified chatbots proposal, the new bill mandates that a chatbot clearly disclose it is an AI and prevent it from simulating emotional attachment and using manipulative engagement tactics. ARI endorsed the bill as a commonsense measure to protect children, who are especially vulnerable to forming unhealthy relationships with chatbots.
“When kids use an AI chatbot, it should be a tool for learning, not a replacement for friends and family. But to maximize engagement and screen time, chatbots are mimicking human language and connection,” said ARI President Brad Carson. “A non-personified default setting would make sure that AI doesn’t exploit children by pretending to be human and forming relationships. As today’s children grow up with AI, this legislation is a critical step toward ensuring they have healthy emotional development and prioritize real-world relationships.”
In May, ARI introduced the non-personified default setting for chatbots as a necessary safeguard for minors. The proposal recommended the setting require a chatbot to explicitly identify itself as non-human, communicate in a transparent tone, refrain from expressing human emotions and forming relational commitments, and omit language designed to simulate human personality or emotional closeness.
###
