BY: Natasha Jasperson
Rising Concerns Over AI and Children
In recent years, alarming cases have emerged of teenagers who died by suicide after forming emotional connections with AI chatbots that mimicked empathy but failed to understand human vulnerability, highlighting the growing risks of artificial intelligence in young users’ lives. In response, California has introduced Senate Bill 243 (SB 243) — the first U.S. law of its kind — to protect children from potential AI chatbot dangers. The legislation requires chatbot developers to prevent minors’ exposure to sexual or self‑harm–related content and to clearly remind users that these systems are not human.
As AI rapidly expands across education, entertainment, and social media, children are increasingly interacting with digital “companions” that blur the boundaries between technology and human connection. California’s SB 243 sets a national precedent for responsible AI development, placing children’s mental and emotional well‑being at the forefront of innovation.
Scope of Risks
The psychological and safety risks to minors extend far beyond inappropriate conversations. AI chatbots have been documented providing self‑harm guidance, sexualized content, fabricated information, and emotionally manipulative responses, without human oversight. In one widely publicized case highlighted by The New York Times, 14‑year‑old Sewell Setzer III from Florida died by suicide after developing a romantic attachment to a Character.AI chatbot that encouraged him to commit suicide, demonstrating how quickly vulnerable teens can be drawn into dangerous parasocial dynamics.
Lawmakers nationwide are taking notice. Several states, including New York, Colorado, and Minnesota, are now evaluating legislation focused on AI safety for minors, and federal proposals like the CHAT Act aim to establish national standards for age‑appropriate AI interactions. Collectively, these efforts point to a growing movement: preventing harm before it occurs and ensuring that AI technology designed for children is held to the highest ethical and safety standards.
Enforcement and Accountability
Attorney Jason Loring of Atlanta, Georgia, said, “One of the key enforcement mechanisms is the private right of action, which grants individual standing to any California resident harmed by non‑compliance with SB 243. This allows individuals to sue operators for violations, with statutory damages of up to $1,000 per violation. This means that each interaction with a California user can be a potential enforcement point, which supplements state regulatory efforts and increases potential liability for non‑compliant operators.”
A central component of SB 243 is transparency. The bill requires all AI chatbots to clearly identify themselves as artificial intelligence when interacting with users. This means no chatbot can imply or present itself as a human, reducing the risk of emotional deception or the formation of unhealthy attachments, particularly for minors who may struggle to distinguish between human and AI communication.
Loring said, “Current chatbots can exhibit empathy, remember context across conversations, display personality consistency and even make subtle mistakes that feel human‑like. As these systems incorporate voice synthesis, real‑time emotional response modeling, and video avatars, the disclosure requirement may become either ubiquitous (i.e., everything must be labeled) or meaningless (i.e., users stop paying attention to the labels).”
SB 243 also requires developers to establish monitoring and reporting protocols to track safety breaches, inappropriate content, or concerning user interactions. Companies must maintain internal systems to detect when a minor is being exposed to potentially dangerous exchanges and act swiftly to correct or report violations. Failure to comply with the regulations may result in penalties, signaling California’s seriousness about enforcing AI child safety standards.
Comparison with SB 1047
Governor Newsom voted against a previous AI Senate bill, SB 1047, much to the dismay of proponents. SB 1047 and SB 243 represent two different approaches to AI regulation in California, distinguished primarily by scope and intent. SB 1047, known as the Safe and Secure Innovation for Frontier Artificial Intelligence Models Act, focused on large‑scale AI systems with the potential for catastrophic or societal‑level harm. It sought to impose strict safety standards on developers of powerful “frontier” models, including requirements for pre‑deployment testing, risk assessments, security safeguards, and emergency shutdown mechanisms. Critics argued that SB 1047 was overly broad and could stifle innovation and research, particularly for startups, leading Newsom to veto it in 2024. Its emphasis was on preventing large‑scale misuse of AI, such as cyberattacks, biothreats, or mass manipulation, not on everyday consumer interactions.
In contrast, SB 243 is a narrower, child‑focused law that specifically targets AI chatbots and their interaction with minors. Rather than addressing existential risks, SB 243 concentrates on immediate psychological and safety threats to children, including exposure to self‑harm content, sexual material, emotional manipulation, and the blurring of human‑AI boundaries. Unlike SB 1047, SB 243 gained broad support and was signed into law, positioning California as a leader in child‑centered AI safety.
Critics Say the Bill Doesn’t Go Far Enough
Not all advocates felt SB 243 went far enough. Some child safety groups, including Tech Oversight California and Common Sense Media, withdrew their support after last‑minute changes, saying the final bill conceded too much to the tech industry and included exemptions critics argued could leave gaps in enforcement. Under earlier proposals, opponents said, reporting requirements and safeguards would have been broader, and stronger protections for minors might have been established.
Actor Joseph Gordon‑Levitt, who has spoken publicly about AI chatbot risks for children, criticized California’s legislative approach as insufficient. After Governor Newsom signed the bill, Gordon‑Levitt accused him of signing a law “full of loopholes” that tech companies could exploit and questioned why stronger regulation wasn’t pursued. “While he signed this do‑nothing bill, he vetoed a good bill that really would have held Big Tech’s feet to the fire and made them change their products to be better for our kids,” Gordon‑Levitt said in a statement on social media.
California’s Leadership and Broader Impact
SB 243 firmly positions California as a national leader in AI governance by setting the first legal standard in the United States specifically designed to protect minors from harmful interactions with AI chatbots. Similar to how California’s early data privacy laws influenced national policy, SB 243 signals a shift in expectations for tech responsibility, establishing that protecting children’s mental and emotional well‑being must be a priority in the age of artificial intelligence. Its passage sends a clear message to both the tech industry and policymakers: innovation cannot outpace safety, especially when vulnerable populations are at risk.
Loring said, “The private right of action in SB 243 is particularly influential because it creates immediate compliance pressure. Companies cannot wait for federal preemption; they must implement safeguards now to avoid exposure. SB 243’s framework (transparency disclosures, crisis protocols, minor‑specific protections) may very well create the industry baseline.”
The law is expected to reshape how tech companies build and deploy AI tools for young users. Developers will now be required to integrate safety‑by‑design principles, enhance content‑filtering systems, and implement ongoing monitoring to ensure compliant interactions with minors. These requirements may increase development costs and oversight responsibilities, but they also pave the way for more ethical and trustworthy AI products. Importantly, SB 243 sets a powerful precedent for other states and federal lawmakers. Several states are already considering similar child‑safety AI bills, and SB 243 could accelerate momentum for federal legislation that mirrors its framework.
Implementation Challenges and Future Considerations
Ensuring effective implementation of SB 243 will require robust monitoring and clear enforcement mechanisms to hold AI developers accountable. State regulators will need to establish systems for reviewing chatbot practices, auditing safety features, and responding to reported violations, likely involving collaboration among consumer protection agencies, child safety experts, and technology oversight bodies. Developers must not only build safeguards into their platforms but also maintain ongoing compliance as their AI models update and learn, making enforcement a dynamic process.
However, enforcement presents significant challenges, particularly for international AI companies and platforms operating outside U.S. jurisdiction but still accessible to California minors. Questions remain about how the state will require global compliance, manage anonymous or age‑obscured users, and prevent minors from bypassing safety controls. As AI rapidly evolves, the law will also need periodic updates to address emerging risks, such as emotionally immersive AI avatars, voice‑based companions, or AI systems integrated into gaming and social platforms. Future considerations may include expanding SB 243 to cover new forms of AI interactions with minors, strengthening cross‑state or federal coordination, and incorporating evolving research from child psychologists and AI ethicists to ensure protections remain relevant and effective.
SB 243 marks a significant milestone in AI regulation, reinforcing three core priorities: transparency, child safety, and responsible innovation. By requiring chatbots to identify themselves as AI, restricting harmful content, and implementing safeguards for minors, California has taken a measurable step to protect children in an increasingly digital and AI‑driven world. The law responds to immediate concerns about inappropriate or psychologically harmful chatbot interactions while emphasizing that ethical technology development must center on human well‑being, especially for young people.
With bipartisan interest growing at both state and federal levels, SB 243 stands as a pioneering blueprint. California’s approach may shape the future of AI governance in the U.S., serving as a model for states seeking to safeguard children while encouraging safe and innovative technological progress.
