California is close to passing strict rules for artificial intelligence tools used by children. Under pressure over youth mental health, tech companies are shifting policies. OpenAI recently announced its support for Senate Bill 1119, which aims to protect minors from harmful interactions with "companion chatbots." This move aligns with the developer's focus on youth safety, including the launch of ChatGPT for Teens.
What Happened
Under California Senate Bill 1119, artificial intelligence companies face strict guardrails when designing tools for minors. According to the official bill text, operators of companion chatbots must actively prevent their systems from posing an "unreasonable risk of a covered harm to a minor." They must conduct child safety risk assessments and submit to independent safety audits starting before July 1, 2027.
To enforce these measures, the legislation shifts the burden of age identification. According to the Angelo Law Firm, platforms must request age bracket data when an application is downloaded. If a company fails to verify ages, it must apply child-safety protections to all users or face civil penalties up to $15,000 per intentional violation. The Sacramento Bee also reports that companion bills aim to limit screen addiction by banning push notifications during school hours and overnight.
In a public letter, OpenAI endorsed the bill. The company noted in its official announcement that the legislation establishes safeguards while preserving access to tools that help students learn.
The Bigger Picture
OpenAI reports that nearly nine in ten teens use ChatGPT weekly for learning and productivity. However, independent data reveals a wider range of habits. According to a study analyzed by Elizabeth Nick, PhD, adolescents interact with up to 88 different generative AI applications. Many of these are companion or roleplay apps, such as "Janitor AI" and "Wemate," which allow intimate, unregulated conversations.
Teens are also using chatbots as primary sources of emotional support. Research published by Education Week indicates that many young people turn to AI for advice after experiencing distressing or confusing offline events, including sexual harassment.
These concerns are driving changes across the tech industry. As we previously reported, OpenAI has expanded its educational footprint while facing growing privacy demands. Other companies are taking similar steps. For example, Google.org hiring new leadership reflects efforts to direct global youth safety initiatives.
What This Means for Families
For parents and educators, California's legislation and OpenAI's response signal a shift toward default protections. According to the company's teen guidelines, OpenAI's "ChatGPT for Teens" automatically places users aged 13 to 17 into a restricted environment. This mode limits romantic engagement and prohibits chatbots from claiming they are human. It also provides parents with monitoring tools.
However, automatic age estimation brings privacy trade-offs. The Electronic Frontier Foundation warns that verifying a user's age online often requires sharing sensitive personal data, which raises questions about how companies retain this information. Legal experts at Baker Donelson also point out that platforms face massive liability if their age-verification systems fail to protect underage users from inappropriate content.
What You Can Do
Parents can take several immediate steps to protect their children. First, talk to your child about the specific AI tools they use, keeping in mind that they may be interacting with roleplay or companion bots rather than standard homework helpers. Next, set up OpenAI's parental tools to monitor settings and receive alerts for high-risk interactions. Finally, discuss the limits of AI relationships and remind your child that AI models are not human, so they cannot offer genuine friendship or professional mental health advice.