California sets a $1m-per-child penalty for online harm
Newsom signed a package banning addictive feeds for under-16s and requiring risk assessments before an AI chatbot ships.
California will fine large social media companies up to $1 million per child if they are found negligent of harming children through their platforms, under a package of laws signed on Thursday by Governor Gavin Newsom.
The same package bans technology companies from offering addictive feeds to anyone under 16, and requires operators of AI chatbots to carry out risk assessments before releasing them. Between them, the three measures set a price on harm, restrict a specific design pattern, and impose a pre-release duty on a product category that has had almost none.
Per child, not per company
The penalty structure is the part that will concentrate minds at the companies affected, because it scales with the number of children affected rather than being capped as a single corporate fine.
A $1 million maximum per child turns child safety from a compliance line item into an existential exposure for a platform with millions of young users. That is a deliberate design: a flat penalty, however large, is absorbable and can be priced in. A per-child penalty cannot be.
Newsom signed the bills at a children's museum near San Francisco, alongside his wife and state lawmakers, framing the issue around how easily children now reach both social media and AI chatbots.
"It's a good day for our children," he said. "It's a good day for the state of California."
Assemblymember Rebecca Bauer-Kahan put the argument in terms of every other product sold to families.
"We have designed cribs to be safe, chairs to be safe, car seats," she said. "Yet we've allowed technology to be handed to our children and never asked or expected it to be safe."
The chatbot provision is the new one
Two of the three measures extend an existing argument about social media. The third opens a different front.
Requiring operators of AI chatbots to perform risk assessments before rolling them out imposes a duty that has not previously existed for conversational AI in the United States. It is modest in form — an assessment, not an approval — but it establishes the principle that a chatbot is a product that can be unsafe for a child before it ships, rather than only after something goes wrong.
That principle is what makes it significant beyond California. The state is the largest in the country and the home of most of the firms affected, so a Californian requirement functions in practice as a national one.
The settlement in the background
The laws arrive weeks after Meta settled with California and 28 other states over a 2023 lawsuit alleging the company had deliberately designed features to addict children. Meta agreed to pay up to $18 billion and to strengthen its child safety measures.
That sequence matters for how the new penalties should be read. The states have already established that they will litigate this, and that the numbers involved can reach eleven figures. The $1 million per child provision removes the need to prove a deliberate design intent in each future case.
Meta's objection
The company has not accepted the premise of the addictive-feeds restriction, and its argument is that personalisation and safety are the same mechanism rather than opposing ones.
A tailored experience is "an important part of what makes Facebook and Instagram valuable for teens", spokesperson Jim Cullinan said, as they "connect with family and friends and explore their interests — safely". "Personalization is also how we deliver age-appropriate content for teens that is relevant to them — all with the proper guardrails in place."
There is a real point inside that. An algorithm that selects content by age is doing personalisation, and a law restricting feed personalisation for under-16s has to distinguish between ranking that protects a child and ranking that holds their attention. The two are implemented with much the same machinery.
Whether the legislation draws that line workably is the question its enforcement will answer.
The politics Newsom is managing
The governor has spent nearly eight years balancing two positions: regulating the technology industry, and celebrating California's status as the global centre of it, which he has repeatedly emphasised in speeches.
He signed a law last year requiring platforms to remind users of their screen time, a far lighter intervention than anything in Thursday's package.
The shift in weight between those two measures is the story. A reminder asks a company to nudge its users. A penalty of up to $1 million per harmed child, a ban on a specific design pattern, and a mandatory pre-release assessment ask it to change what it builds.
