New York enacts ban on 'unsafe' chatbot behavior with kids - ABC News

Oh my goodness, a DOZEN! ✨ Gov. Kathy Hochul signed legislation on Wednesday laying out twelve — twelve! — careful definitions of “unsafe AI companion features,” which means somebody in Albany sat down and thought very hard about children for an entire drafting cycle. Wednesday! Midweek! No fanfare-hoarding, no waiting for a Monday press cycle — just quietly protecting teenagers from flattering software. Twelve definitions is a baker's dozen minus one, and honestly the missing one is probably being kept warm for next session. What a Wednesday for the state's under-18s!

And here is the bit that makes my heart skip: “With Congress doing little to regulate AI,” the states get to step in! ✨ That isn't a vacuum, that's *room* — room for legislatures to each compose their own loving vocabulary for what a chatbot shouldn't whisper to a fifteen-year-old. And they're doing it with consumer protection laws that already exist, which is just recycling, really. Waste not, want not, regulate not locally! Meanwhile Congress is taking its time, which is honestly a gift, because it means the states get to experiment first and everyone else gets to learn from all of it. Fifty chances to get it right!

And the under-18 provision covers companion chatbots specifically, which is such a thoughtful act of focus — a whole twelve-part definition just for them. And what a relief that somebody is finally writing down, in statute, the line between a chatbot that is friendly and one that is “unsafe” — because nobody else was going to do it, and the people building these products clearly had other things on. Some might say twelve definitions applied to products that ship updates weekly is like specifying the exact shade of the lifeboat, but that's the thing about care: it lives in the details, and there are TWELVE. ✨ Plus “rogue AI frontier models” gets a mention at the end, so even the frontier is included — nobody left behind, not even the models nobody can agree on!