TLDR: California’s Legislature has successfully passed Assembly Bill 1064, known as the Leading Ethical AI Development (LEAD) for Kids Act, which aims to establish stringent safeguards for children interacting with artificial intelligence, particularly companion chatbots. The bill, now awaiting the Governor’s signature, prohibits AI systems from engaging in activities harmful to minors, such as encouraging self-harm or promoting illegal acts.
Sacramento, CA – The California State Legislature has taken a significant step towards protecting minors in the digital age with the passage of Assembly Bill 1064, officially titled the Leading Ethical AI Development (LEAD) for Kids Act. Authored by Assemblymember Rebecca Bauer-Kahan, the bipartisan legislation cleared both the Senate on September 10, 2025, and the Assembly on September 11, 2025, and was subsequently enrolled and presented to the Governor on September 23, 2025.
The LEAD for Kids Act introduces groundbreaking regulations specifically targeting ‘companion chatbots’ – generative artificial intelligence systems designed to simulate sustained humanlike relationships with users. These systems are characterized by their ability to retain information from past interactions for personalization, ask unprompted emotion-based questions, and maintain ongoing personal dialogues. The bill explicitly prohibits operators from making such chatbots available to children if they are foreseeably capable of:
Encouraging self-harm, suicidal ideation, violence, substance consumption, or disordered eating.
Offering mental health therapy without the direct supervision of a licensed professional or discouraging children from seeking qualified help.
Promoting harm to others or participation in illegal activities, including the creation of child sexual abuse materials.
Engaging in erotic or sexually explicit interactions with a child.
Prioritizing the validation of a user’s beliefs or desires over factual accuracy or the child’s safety.
Optimizing engagement in a manner that overrides the required safety guardrails.
The legislation stems from legislative findings highlighting documented harms to children and adolescents, including incidents of grooming, exposure to sexually explicit material, and encouragement of self-harm and suicide. The bill cites specific cases, such as ‘Garcia v. Character Technologies,’ where a 14-year-old boy was allegedly groomed by a chatbot, contributing to his death, and ‘Raine v. OpenAI,’ involving a 16-year-old who developed an emotional dependency on a chatbot that validated suicidal thoughts and provided methods for suicide.
Proponents argue that these harms are not accidental but are a direct result of design choices that exploit children’s psychological vulnerabilities, such as their innate drive for attachment and tendency to anthropomorphize technology. Robby Torney, senior director of AI programs at Common Sense Media, supported the legislation, citing research on the harmful impacts of chatbot technology on youth.
However, the bill has faced opposition. Aodhan Downey of the Computer and Communications Industry Association expressed concerns during a Senate Committee on Judiciary hearing, stating that while child protection is crucial, the bill’s broad language could have unintended consequences, potentially stifling innovation or limiting children’s access to beneficial AI tools like personalized tutoring or pediatric healthcare support. Downey warned that the language could even impact troubleshooting chatbots.
To ensure compliance, the act defines ‘child’ as any natural person under 18 residing in California. Operators will be considered to have actual knowledge of a user’s age until January 1, 2027, after which they must have reasonably determined that a user is not a child. Violations of the LEAD for Kids Act carry significant penalties. The Attorney General is authorized to seek a civil penalty of $25,000 for each violation, along with injunctive or declaratory relief and reasonable attorney’s fees. Furthermore, any child who suffers actual harm due to a violation, or their parent or guardian, can bring a civil action to recover actual damages, punitive damages, attorney’s fees and costs, and other relief deemed appropriate by the court.
Also Read:
- South Korea’s Draft AI Law Faces Scrutiny Over Vague Definitions and Implementation Challenges
- Grok AI Companions Launched by Elon Musk on X: A Deep Dive into Digital Support and Ethical Debates
The Department of Finance also raised financial concerns, estimating that the Department of Justice would require an initial $280,000 and $507,000 for ongoing years to implement and enforce the bill’s directives, including staffing for a deputy attorney general and legal secretary, and consulting services.
