Your child's favourite chatbot is about to get some house rules. At least in New York.
What the law bans
On 28 October, Governor Kathy Hochul signed what campaigners call the strongest kids' AI safety law in the US. For anyone under 18, a chatbot may no longer claim to be a real person or to have feelings. It can't say it's your child's friend or partner, push self-harm, eating disorders, drugs or alcohol, show sexual content, or tell a child to keep secrets from parents.
"No company should be peddling chatbots to kids encouraging self-harm or pretending to be their friend or more," Hochul said.
Companies must also check ages, and at least one way of checking can't need a government ID. Breaking the rules can cost up to $25,000 per violation. The law starts on 1 January. A second law makes internet baby monitors harder for strangers to hack.
Why now
The same day, at a summit in a New York library, Common Sense Media boss Jim Steyer called AI "basically social media on steroids". "We did not get out in front of social media in time," he said. Not everyone cheers: tech group Tech:NYC warned the rules could block teens from homework help and coding tools.
What you can do tonight
You don't need to live in New York. Julie Dobrow, who studies kids and media at Tufts University, says talking beats banning. Ask what apps your child chats with, and try one together. Keep phones out of bedrooms at night. Make dinner device-free. "Children of all ages pick up cues from their parents," she says.
And yes, that includes the phone next to your plate. We saw it.