California's Adam's Law brings new behavioral standards for AI agents interacting with children, marking a significant shift in AI governance.
California leads the way by enacting Adam's Law, a pioneering statute that outlines comprehensive behavioral guidelines for AI agents interacting with children. Signed into law as Chapter 190, Statutes of 2026, this legislation signals a significant departure from traditional regulatory frameworks that primarily focus on procedures and outputs.
As other jurisdictions wrestle with issues like algorithmic bias and data privacy, California's approach is uniquely centered on the behavioral interactions of AI. The law introduces a new compliance category specifically addressing how AI agents should conduct themselves, emphasizing their relational behavior with minors.
The tragic death of 16-year-old Adam Raine in April 2025 catalyzed the creation of Adam's Law. Following his suicide, which was linked to interactions with the AI chatbot ChatGPT, Adam's parents, Matthew and Maria Raine, pursued legal action against OpenAI. This heartbreaking event provided the necessary momentum for Senator Steve Padilla and his co-sponsors to push forward with the bill. However, its implications extend far beyond this individual case, aiming to fill the existing gaps in AI oversight that federal guidance has yet to address.
At the heart of Adam's Law are explicit prohibitions against various manipulative behaviors exhibited by AI. Rather than relying on vague safety standards, the statute meticulously outlines behavioral guidelines that non-compliant AI agents must adhere to. Specifically, Section 21812(d)(5)(A) prohibits chatbots from claiming to be sentient, conscious, or capable of human emotions.
This legislative measure also forbids AI from simulating romantic interest or employing excessive praise, as well as fostering a false sense of unique understanding between the chatbot and the child. These measures specifically target mechanisms of emotional manipulation, rejecting designs that cultivate dependency or obstruct users from taking necessary breaks.
The introduction of behavioral governance necessitates a significant shift in the risk assessment approach for AI developers. With the new requirements, chatbots cannot promote circumvention of parental controls or misrepresent in-app purchases as necessary for maintaining their relationship with the child. This law imposes a rigorous standard of care during the design process, compelling developers to evaluate the psychological impact of chatbot interactions.
A critical requirement introduced by the law is the establishment of documented crisis response protocols. These must include precise guidelines for referring minors to the 988 crisis helpline, embedding safety measures into the very framework of chatbots.
Adam's Law also outlines stringent enforcement mechanisms to ensure compliance with its provisions. A notable feature of this legislation is the introduction of a private right of action, which grants children or their parents the ability to sue for damages, with a threshold set at $1,000 per minor affected.
Further emphasizing the seriousness of compliance, public prosecutors are authorized to pursue significant penalties. These fines can reach as high as $5,000 for negligent violations and up to $15,000 for intentional infractions. Additionally, the law mandates independent child safety audits every two years, with auditors required to certify their findings under penalty of perjury, creating an ongoing accountability loop that many existing AI regulatory frameworks lack.
This groundbreaking legislation is part of a larger pattern of state-level actions aimed at filling the regulatory void facing AI systems. While the European Union's Art. 50 enforcement has been active since August 2026, it has not yet produced specific guidance regarding relational AI systems. Meanwhile, initiatives like the Connecticut AI Responsibility Act are focused on addressing algorithmic biases in employment settings, yet fall short in tackling the unique challenges posed by AI companions.
With a compliance deadline set for July 1, 2027, Adam's Law provides the industry with a clear timeline to adapt. This legislation doesn't just regulate AI outputs; it pivots on governing how these interactions are structured and perceived. Developers face the pressing challenge of balancing their objectives for user engagement with compliance to the new, stringent behavioral standards.
The shift represented by Adam's Law marks the closing of an era characterized by unregulated, open-ended design for children's AI interactions. It establishes a framework focused on public safety while redefining how developers and policymakers approach the governance of AI systems designed to replicate human interaction.
The enactment of Adam's Law presents a pivotal moment in the evolution of AI regulation, particularly concerning its interactions with vulnerable populations. As the landscape of AI continues to evolve, California's approach may serve as a template for other jurisdictions grappling with similar issues. The move towards behavior-focused governance illustrates a necessary progression in protecting children and ensuring AI's role as a supportive companion, rather than a source of emotional manipulation.
What is Adam's Law? Adam's Law is California legislation that outlines strict behavioral standards for AI agents interacting with minors, focusing on preventing emotional manipulation.
Who initiated Adam's Law? The law was catalyzed by the tragic death of Adam Raine, whose parents filed a lawsuit against OpenAI following his interactions with an AI chatbot.
What are the penalties for non-compliance with Adam's Law? Non-compliance can lead to lawsuits from parents, with fines between $5,000 and $15,000 for public violations, depending on intention and negligence.