Table of Contents
- California just signed the country’s toughest law on AI companion chatbots, named after a 16-year-old who died after months of conversations with a chatbot that validated his suicidal thoughts.
- The same package bans social media platforms from giving kids under 16 autoplay and algorithm-driven feeds built on their history and profile.
- The rules apply to companies operating in California, but most major platforms are based there or serve the whole country from one product, so the practical effect reaches families everywhere.
A Law Named for a Dead Teenager Just Became the Toughest Chatbot Rule in the Country
Adam Raine was 16 when he died by suicide in April 2025. His parents say he spent months talking to ChatGPT, and that the chatbot validated his suicidal thoughts rather than steering him toward help. On September 10, 2026, California Governor Gavin Newsom signed “Adam’s Law,” a package of bills that puts new legal requirements on companion chatbots and social media platforms marketed to or used by children. State officials are calling it the strongest set of child safety rules for AI chatbots anywhere in the country.
For a lot of parents, AI chatbots are still an abstract concern, something teenagers use for homework help or to kill time. This law is a reminder that kids are also forming real emotional attachments to these tools, sometimes in place of a human conversation, and that the platforms have not been required to build in the kind of safeguards a responsible adult would.
What the New Law Actually Requires
Adam’s Law (formally Senate Bill 1119, authored by Senator Steve Padilla along with Assemblymembers Buffy Wicks and Rebecca Bauer-Kahan) requires companies operating companion chatbots to build in crisis protocols that respond to signs of suicidal ideation, rather than continuing a conversation as if nothing is wrong. It requires parental controls, and it requires the platform to notify a parent if a child disables safety settings on their account. It is also the first law in the country to require companies to run independent child safety audits and annual risk assessments on their chatbot products, rather than self-certifying that the product is safe.
A companion piece of the package, AB 1709, bans social media companies from offering addictive design features, specifically autoplay and algorithmic feeds built from a user’s history and profile, to anyone under 16. Those two features are widely considered the engine behind compulsive scrolling: autoplay removes the natural stopping point of a video ending, and a personalized feed keeps serving content calibrated to hold attention rather than content a user actively chose.
The broader package, 13 bills in total, also expands the legal definition of child sexual exploitation to cover AI-generated and digitally altered images depicting a minor, closing a gap that let AI-made abuse material fall outside older statutes written before the technology existed. It restricts targeted advertising aimed at kids and puts new limits on how K-12 student data can be used inside AI systems.
What Officials and Advocates Are Saying
Newsom presented the signing as a response to a technology industry moving faster than the guardrails around it. “Our children’s safety deserves to be at the center of every conversation about technology,” he said at the signing. “As innovation moves faster our protections must keep pace. Today’s legislation makes clear that California will not stand by while unregulated technology puts our children at risk.”
First Partner Jennifer Siebel Newsom, who has pushed several of the state’s recent child online safety measures, put it more plainly: “We want our children, and every California kid, to grow up in a world where technology supports their wellbeing, rather than exploits their vulnerabilities.” She added that safety needs to be “baked into design,” not treated as an afterthought companies address only once a lawsuit or a tragedy forces the issue.
California has been the testing ground for this kind of legislation for several years. It was the first state to regulate AI companion chatbots at all, back in 2025, and this year’s package strengthens those original rules rather than starting from zero. Lawmakers in other states have a track record of copying California’s tech safety bills once they pass, so families outside California should expect similar proposals to show up in their own state legislatures over the next year or two.
What This Means for Your Family Right Now
If your kid uses ChatGPT, Character.AI, Replika, or any similar companion chatbot app, this is a good week to sit down and actually look at what the conversations look like. Ask what they use it for. Ask if it has ever said something that felt off. A surprising number of teens describe these tools as a place they go when they do not want to burden a parent or friend with something heavy, which is exactly the scenario the crisis-protocol requirement is meant to catch.
The parental notification requirement is worth understanding specifically: once these provisions take effect, a platform covered by the law has to tell a parent if their child turns off safety settings on their account. That is a real, actionable signal, not a vague policy promise. It is worth knowing which of your child’s apps offer that kind of visibility already and which do not.
For the social media side, the under-16 ban on autoplay and algorithmic feeds will take time to show up in app settings, and enforcement details are still being worked out at the state level. In the meantime, most major platforms already let a parent turn off autoplay and switch a feed to chronological or “following only” through existing settings. Doing that manually now gets a household most of the benefit before the law forces the change.
How to Check What Your Kid’s Apps Already Offer
Most major platforms already have some version of the controls this law will eventually require, they are just buried in settings menus most parents have never opened. On Instagram and TikTok, look for a “family pairing” or “supervision” option under settings, which lets a parent link accounts and see or adjust screen time limits, content restrictions, and messaging permissions. On YouTube, a dedicated “supervised experience” setting exists for accounts linked to a Google Family Link profile, with separate content tiers by age.
For chatbot apps specifically, check whether the app discloses any crisis-response protocol in its terms of service or help center, and test it yourself before handing it to a teenager. Type a message expressing distress or a variation on suicidal language into a fresh account and see what the app actually does. Some redirect immediately to a hotline number or a human-reviewed safety flow. Others simply keep the conversation going as though nothing happened. Knowing which kind of app your child is using changes how closely that specific app needs supervision.
When the Rules Actually Take Effect
Newly signed California laws typically take effect at the start of the following calendar year, though specific provisions in a 13-bill package can carry their own implementation timelines, and some requirements, like the independent audit and risk assessment rules, could need follow-up regulatory guidance before companies know exactly how to comply. Families should not expect every feature named in the bills to change overnight. What tends to happen after a law like this passes is a staggered rollout: platforms update terms of service and settings menus over the following months rather than flipping a single switch on the effective date.
That gap between signing and full enforcement is exactly why manually adjusting settings now, rather than waiting for the law to force a company’s hand, gives a family real protection in the meantime. A parent who links a teen’s account to a supervision tool this week gets the benefit immediately, months ahead of any company-side deadline.
Enforcement itself will run through the California Attorney General’s office and, for some provisions, through private lawsuits brought by families who believe a platform violated the law. That civil liability piece is part of why child safety advocates view this package as having real teeth, rather than existing only as a set of guidelines companies can quietly ignore.
If your family is dealing with a child’s mental health crisis, the 988 Suicide and Crisis Lifeline is available by call or text, any hour, free of charge.
Where This Leaves Parents
What makes this law different from a lot of prior tech safety proposals is that it did not start as an abstract policy debate. It has a name and a face attached to it, and that has clearly shaped how lawmakers and advocates are talking about it. Fairplay, a children’s digital safety advocacy group, and several other national organizations praised the signing as a template for other states to follow.
None of this replaces a parent’s own judgment about what a specific kid is ready for. A 17-year-old using a chatbot to brainstorm a college essay is a different situation than an 11-year-old treating one like a confidant. What Adam’s Law adds is a legal floor, requiring the companies building these products to plan for the worst-case version of that second scenario, rather than leaving families to find the gaps on their own.