Children are increasingly turning to AI chatbots for advice, mental help and companionship, experts say. Lawmakers are worried that behavior can be more harmful than helpful if guardrails aren’t put into place.
Sen. John Curtis introduced a bipartisan bill on Tuesday to establish a federal framework regulating online chatbots, particularly as it relates to usage by teenagers and young children. The Safeguarding AI Features to Ensure Kids’ Informed Digital Safety Act, or SAFE KIDS Act, would implement a number of requirements for chatbots before they can go live.
“Parents deserve confidence that AI tools are not exposing their children to harmful content, fostering unhealthy emotional dependence, or exploiting their personal information,” Curtis said in a statement. “We need clear standards for transparency, accountability, and child safety so families can navigate the opportunities and risks of emerging technologies with confidence.”
AI chatbots are online generators that mimic conversations with human beings. The bots can be used for educational and research purposes, but they are increasingly being used as a type of online companion.
What would SAFE KIDS Act do?
The SAFE KIDS Act would require chatbot providers to conduct “rigorous, ongoing risk assessments” before making these systems available for download or usage. It would also ban advertising targeted toward kids or the sharing of a child’s personal data without parental consent.
The bill would also ban AI chatbots from mimicking human emotions; ban bots from generating sexual images; require providers to provide crisis resources and notify parents if a child is at risk of causing harm, such as suicidal ideation or self-harm; and encourage providers to use age estimation technology.
Chatbot providers would then have to implement annual safety audits to ensure they are in compliance with the law, and make those findings available to the public.
“AI chatbots that promote companionship pose significant risks to young and developing minds,” California Sen. Adam Schiff, who co-introduced the legislation, said in a statement. “These chatbots can represent the power of the entire internet in a humanlike form, and impressionable kids need to be protected from its worst impulses — and from seeing their data misused by the companies behind them.”
The bill comes as lawmakers in both parties are searching for ways to protect children online. The Senate is set to consider a number of bills seeking regulations for social media companies and AI chatbots, and the House is moving forward on a bipartisan package as well.
The proposal builds on legislation from Sen. Ted Cruz, R-Texas, known as the CHATBOT Act that would require parents to create accounts for children younger than 13 years old, and it would allow parents to set the controls, among other things. Curtis is a co-sponsor of that bill.
Other bills, such as the Kids Online Safety Act, push to require social media apps to “exercise reasonable care” in their design and features that typically cause teenagers to become addicted to the online sites. Lawmakers are also exploring ways to require companies to meet certain standards before implementing algorithms that “select, order, or prioritize information presented to users based on user-specific data.”