You know the scene. Five more minutes turns into an hour, the app keeps finding a reason to pull your child back, and stopping feels like a fight most nights. On September 10, California answered a version of that worry, and Governor Newsom signed what his office called the nation’s strongest child-safety laws for chatbots and social media. The protections are real: crisis help, disclosure, parental notification, liability when a product harms a child. But the centerpiece chatbot law governs what happens after a harmful moment, not the engineered design that makes the product so hard to put down in the first place, and that gap is the whole story.
California has signed the country’s strongest child-safety tech laws, and the protections are real. But the companion-chatbot bill governs what happens after a harmful moment, not the engineered design that makes the product so hard to put down. Here is what that leaves for a parent to carry.
California’s new laws have parents asking what the rules actually do and how to judge the apps and chatbots their kids already use. Here are the questions coming up most.
Common questions
What did California sign on September 10, 2026?
Does Adam’s Law make companion chatbots safe for my child?
How do I tell if an app or chatbot is designed to be hard to put down?
My child is already pulled to these apps and struggling. Where do I start?
California passed the nation's strongest child-safety chatbot law. It covers crisis, disclosure and liability. The one thing it leaves mostly alone is the design built to keep your kid coming back.
What California signed
On September 10, in Marin County, Governor Gavin Newsom signed a package of 13 child-safety and technology bills that his office called the strongest of their kind in the nation. “Our children’s safety deserves to be at the center of every conversation about technology,” Newsom said. First Partner Jennifer Siebel Newsom put the ambition in a single line: “With this critical legislation, California is sending a strong message that when it comes to our kids, safety must not be an afterthought — but baked into design and a necessary requisite for all innovation.”
The centerpiece is SB 1119, known as Adam’s Law, named for Adam Raine, a teenager whose parents, Matt and Maria Raine, stood with lawmakers at the signing. It sets real rules for companion chatbots: age checks, risk assessments before a new chatbot is released, in-app crisis support with a notification to a parent when there is a credible threat of imminent self-harm, parent-controlled default settings, independent safety audits, liability for harmful outputs including self-harm coaching and romantic roleplaying with a minor, and a private right of action. “Adam’s Law sets a new national standard for what we owe our children online,” said Assemblymember Rebecca Bauer-Kahan, a coauthor. It builds on SB 243, the companion-chatbot disclosure law California passed in 2025.
A second bill in the package, AB 1709, reaches somewhere Adam’s Law does not. It bars platforms from offering certain addictive features to users younger than 16 and, in its own words, targets “algorithmically driven feeds and other mechanisms that can encourage prolonged use,” and it sets up an e-Safety Advisory Commission inside the state Department of Justice. “Parents cannot out-parent an algorithm built in a lab to keep kids scrolling,” said Jim Steyer of Common Sense Media.
Parents cannot out-parent an algorithm built in a lab to keep kids scrolling.
Laura Lurns · Learning Success expert
The design the law mostly left alone
Here is the distinction the coverage keeps sliding past. Almost everything in Adam’s Law acts after a harmful interaction has already begun, or after a crisis signal appears: disclose that the bot is not human, refer a child in distress to help, notify a parent, audit the system, carry liability when the output causes harm. Those are substantive protections, and they are worth having. What none of them touch is the thing that makes a companion chatbot compelling to return to in the first place, a product engineered to hold attention. The one bill in the package that names design, AB 1709, aims at social-media feeds, not at chatbots.
The science on that design is clearer than the science on the harm, and the honest version keeps the two apart. Platforms are built around a variable, unpredictable reward, delivered by infinite scroll, engagement-optimized feeds, and notifications that arrive on no schedule you chose. Clark and Zack (2023) describe this as the same addiction-design framework that runs through gambling and gaming: a reward that lands after an unpredictable number of actions produces the most persistent return rate. That is a description of how the products are engineered, stated by the people who have built and studied them, and here is the sharpest limit worth saying out loud: no study we have ties a companion chatbot’s engagement design specifically to a measured harm. The argument that a chatbot is the next, more personal delivery system for that same engineered pull is our reading of how these products are built, not a traced research finding, and we flag it as ours.
Where the harm research does point, it points one direction. The most-replicated social-media-specific harm is upward comparison, measuring yourself against everyone else’s highlight reel, which tracks with lower self-esteem and poorer body image. Vogel and colleagues found it in 2014, and a 2023 meta-analysis pulling together dozens of studies landed in the same place. It falls hardest on the children already watching how they measure up, which is to say the kids already fighting for their own confidence. Two honest caveats travel with that: those studies sit mostly in older teens and adults, not a randomized trial in 13-year-olds, and across all children on average the link between use and well-being is small and tangled up with other things. The strong part of the story is the engineered design and the children most at risk, never a giant average effect on everyone.
Key takeaways
- The gap: Adam's Law handles crisis and liability, not the design that keeps a child coming back.
- Wrong target: The one bill that names feed design aims at social media, not chatbots.
- The real hook: Platforms are engineered around unpredictable rewards, and the design, not screen minutes, holds a child.
What this leaves for you to carry
Start with what the law does for you, because it is not nothing. If your child confides something frightening to a chatbot, there is now a duty to route them to help and, in a credible self-harm emergency, to tell you. A company that builds a bot to coach self-harm or to romance a minor now carries liability for it. Those are real, and a year ago they did not exist. AB 1709 goes further and aims at feed design itself, though it is stepping onto contested ground: a federal judge has already blocked parts of California’s earlier addictive-feeds law, SB 976, while the challenge proceeds, so whether a design restriction survives is unresolved.
What the law does not do is decide, for you, whether a product was built to be hard to put down. That judgment stays at your kitchen table, and it is a better question than screen minutes alone. Watch the design, not the clock. A product that interrupts with notifications you did not ask for, rewards streaks, remembers enough to pull the next session longer, and offers no natural place to stop is engineered to hold attention, whatever its content rating says. The children it holds hardest are the ones already unsure of themselves, and those are the children worth protecting first.
So treat the new law the way you would treat California’s dyslexia-screening push: a real step that helps or falls short depending on how it is used, not a reason to hand the decision off. The protections buy accountability after something goes wrong. The design that makes the product compelling in the first place is still yours to notice, to question, and to build an alternative to. A child who gets the genuine satisfaction of getting better at something real is far harder for an engineered feed to capture.
We cannot make the same mistakes that were made with social media. California is filling the void created by Washington inaction.
Laura Lurns · Learning Success expert
You do not need a state commission to tell you whether a product is serving your child or holding them. You are the one who sees the difference between a child building something real and a child being kept busy, and that is the judgment this whole debate keeps trying to outsource. The villain here is not the California lawmakers, who are wrestling with a genuinely hard problem in good faith, and it is not any one app. It is the design model underneath all of it: building a product around an unpredictable reward so a child keeps coming back, and calling that engagement the same thing as worth. Learning Success was built to refuse that trade. We said no to points, badges and streaks, and built on the one thing no feed is able to manufacture, a child’s relationship with the adult who loves them. A points system makes a child feel like they are learning; real mastery makes them feel capable. Our All-Access membership opens an assessment that asks about the processing systems your child’s learning runs on, then a roadmap that names which skill to build first. It is a place to start building the real thing today, instead of trying to out-willpower an engineered feed.
See what All Access gives your child
See MTSS for schools
Is your child struggling in school?
Get your free personalized learning roadmap
You describe what you see at home. We turn it into a plan you start this week.
- Answer 5 short questionnaires about what you already notice, 30–45 minutes at your own kitchen table
- Your child sits no test and gets no score: nothing to schedule, nothing for them to dread
- You do the answering, the AI does the writing, and a person reviews it before it reaches you
- Access all 40+ courses instantly: reading, math, focus, processing and more, with new ones added regularly
Why we use AI, plainly: it writes from a knowledge base our team maintains and audits. We work through it line by line and pull anything the evidence stops supporting. The roadmap you get on Tuesday reflects what we corrected on Monday, and a human still reads it before you do.
Your school district must evaluate your child free of charge if you ask in writing, whatever your income and whatever the outcome (US, 34 CFR 300.111 and 300.301(b)). That route takes time and answers a different question than you do. This one starts today, from what you already know.
Your answers stay yours. We do not sell your personal information, and we do not hand identifiable assessment data to outside AI companies to train their models.
A screener is a starting point, not a diagnosis. If your child might need formal accommodations (an IEP or 504 plan), or you suspect a vision, hearing or medical cause, pursue a professional evaluation too. That is the only route to those supports.
References
- Office of Governor Gavin Newsom — Governor Newsom signs the strongest child safety chatbot and social media laws in the nation
- Senator Steve Padilla's office — Governor Newsom Signs Adam's Law
- KQED — Chatbots, Data Centers and Surveillance: 5 Silicon Valley Bills Land on Newsom's Desk
- Future of Privacy Forum — Understanding the New Wave of Chatbot Legislation: California SB 243 and Beyond
- Jones Walker LLP — California's SB 243 Mandates Companion AI Safety and Accountability
