'Our Children's AI Safety Package Act' introduced in National Assembly
Bill strengthens child-specific safety design and public protection framework
The National Assembly is moving to build a legal safety net to protect children and teenagers from harmful conversations and dangerous content generated by AI tools such as ChatGPT.
Rep. Jo In-cheol of the Democratic Party of Korea, who represents Gwangju's Seo-gu Gap district, said Monday he had introduced the "Our Children's AI Safety Package Act," a set of bills designed to strengthen the state's responsibility and AI service providers' obligations to ensure children and teenagers can use AI in a safe environment.
The package consists of two bills: an amendment to the AI Basic Act that would incorporate the protection of children and teenagers into national AI policy, and an amendment to the Act on Promotion of Information and Communications Network Utilization and Information Protection that would tighten safeguards for minors using conversational and generative AI services, including chatbots.
According to Jo, generative AI chatbots have evolved well beyond simple search or study tools — children and teenagers now use them as conversation partners, outlets for personal worries and sources of emotional support. A survey by Green Umbrella, a child welfare organization, found that 94.4 percent of children and teenagers said they use generative AI chatbots, and 49.5 percent said they feel understood by AI.
The problem, Jo said, is that institutional safeguards to protect minors from dangerous conversations and harmful content have not kept pace with the rapid normalization of AI use among young people.
The AI Basic Act amendment would classify children and teenagers as a "vulnerable group" in the context of AI and require the government's national AI master plan to address harm prevention and protection for such groups, establishing a state-level responsibility framework for protecting minors in the AI era.
The centerpiece of the information and communications network amendment is imposing child protection obligations on providers of conversational and generative AI services. Specific requirements would include verifying users' ages and identities; obtaining consent from a legal guardian before minors use the service; restricting usage methods and time at the request of a minor or guardian; providing records of conversations — including timestamps and content — to users; and building a reporting system for exposure to risks such as suicide, self-harm, drug abuse and sexual exploitation, along with mandatory measures to remove such content.
"Children today live in an era where they form relationships, share their worries and seek information not only from people but also from AI," Jo said. "As AI has become deeply embedded in children's daily lives, their developmental needs and safety must be the top priority from the very design stage of a service. Protecting children's rights and safety in the digital environment is a challenge of our time that cannot be delayed — I will do everything in my power to see this legislation passed."
bigroot@heraldcorp.com
