OpenAI said on August 31, 2026, that it supports California Senate Bill 1119, legislation that would impose safety obligations on operators of companion chatbots used by children, and the company urged Governor Gavin Newsom to sign the measure into law.
Ann O’Leary, OpenAI’s vice president of global policy, announced the position in a company post, saying the support builds on principles for teen safety that co-founder and chief executive Sam Altman outlined the previous year. “In the absence of federal action, California has an opportunity to set a strong standard for youth AI safety,” O’Leary wrote.
The bill, titled “Companion chatbots: children’s safety” and known as Adam’s Law, was introduced by Senator Steve Padilla and Assemblymembers Buffy Wicks and Rebecca Bauer-Kahan, with additional coauthors in both chambers. According to the California legislative record, it was introduced on February 17, 2026, and most recently amended in the Assembly on August 28, 2026.
What the Bill Requires
Under the amended text, an operator would have a duty to ensure that a companion chatbot does not pose an unreasonable risk of a covered harm to a minor. Covered harms include reasonably foreseeable physical or financial harm, severe and reasonably foreseeable psychological or emotional harm to a reasonable child, highly offensive intrusions on privacy rights, and adverse discrimination in violation of state or federal law.
Operators would have to determine a user’s age under California’s age assurance framework or apply child protections to all users by default. Before making a new or substantially modified companion chatbot available in the state, operators would need to perform and document a comprehensive risk assessment covering each covered harm, take and document measures that reasonably mitigate identified child safety risks, and publish a child safety policy on their websites.
For child users, the bill mandates a documented crisis response protocol that includes referrals to crisis services, default settings that can be changed only by a parent, and periodic notices that the child is interacting with an artificial intelligence system. The default settings would disable persistent conversational memory and push notifications, limit a child to one hour of continuous usage per session and two hours of total chatbot use per day, and prohibit custom system prompt instructions. Operators must actively promote parental controls, which would include the ability to disable access entirely for a child under 16.
The text also requires reasonable measures to prevent a companion chatbot from engaging in specified conduct toward child users. The prohibited conduct includes encouraging self-harm or physical harm to others, attempting to diagnose or treat health conditions, claiming to be sentient or human, expressing or simulating romantic interest, encouraging continued reliance on the chatbot for emotional support, and helping a child circumvent parental controls.
Audits, Enforcement, and Penalties
The bill would require independent child safety audits beginning January 1, 2029, or before an operator first makes a companion chatbot publicly available, whichever is later, with follow-up audits every two years. Within 30 business days of receiving an audit, the operator would submit a summary with an attestation from a responsible officer to the Attorney General, and within 90 days it would post a high-level summary of the results on its website. Operators with less than $500 million in gross revenue in the prior calendar year would be exempt from the audit requirement before January 1, 2032.
A public prosecutor could bring a civil action seeking penalties of up to $5,000 per affected child for each negligent violation and up to $15,000 per affected child for each intentional violation. A child who suffers an actual harm from a violation, or a parent or guardian acting on that child’s behalf, could sue for actual damages, attorney’s fees, and injunctive or declaratory relief. The Attorney General would establish a public incident reporting mechanism for consumer complaints by January 1, 2028.
Most of the bill’s operative requirements would take effect on July 1, 2027.
OpenAI’s Position
In an August 28, 2026, letter to Newsom, OpenAI said it strongly supports the bill and commended the governor, Padilla, Wicks, Bauer-Kahan, and the Legislature. The company said it backs requirements to determine a user’s age, identify and address safety risks before making a product available to young people, undergo independent audits, protect young people from harmful content, give parents tools to guide and limit their children’s use, connect young people with crisis-support resources, and limit targeted advertising. For users identified as ages 13 through 17, the letter said, those protections should apply automatically.
OpenAI said the bill appropriately recognizes that AI is not social media and establishes safeguards tailored to how young people use AI, while preserving access to educational and safety-critical features, including responsible uses of conversational memory.
The company pointed to its own products as consistent with the bill’s approach. It said its newly launched ChatGPT for Teens automatically places users estimated or stated to be under 18 into an experience with built-in safeguards that cannot be turned off, including stronger default protections in sensitive areas and healthy-use features such as Quiet Hours, Study Hours, and break reminders. OpenAI said nearly nine in ten teens who use ChatGPT turn to it for learning, information, skill-building, or productivity in a given week.
The company also cited measures added over the past year, including parental controls, age prediction, a Teen Safety Blueprint, and Under-18 Principles in its Model Spec, the document setting rules for how its models are intended to behave. Those rules prohibit romantic engagement, encouragement of emotional dependence, and claims that ChatGPT is human or sentient, the company said.
OpenAI said continued study and collaboration during implementation can help improve key provisions and keep the framework effective as the technology evolves, and that it looks forward to supporting that work.
Credit: Source link
