Sam Altman thinks ChatGPT should make your kids’ morning podcast. Just talk to them instead.

Sam Altman posted something on Friday that he clearly thought was a winner. A “cool use case,” he called it. Parents could hook up their family calendars, feed ChatGPT Work some context about their kids’ interests, and then every morning on the drive to school, the AI would generate a custom podcast covering that afternoon’s soccer game, an upcoming birthday, some news, whatever.

It’s the kind of vision that sounds great in a keynote and falls apart the moment you actually have a kid in the backseat asking why the robot knows about their birthday. Alex Hirsch, creator of Gravity Falls, summed it up in nine words: “What if you just talked to your children?” That response has been liked over 122,000 times, compared to Altman’s original post’s 9,600 likes. The ratio tells you everything about how this landed.

This isn’t new territory for tech CEOs. Satya Nadella said last year he’d replaced his favorite podcasts with an AI chatbot that could summarize them for him during his commute. Which is a choice. Altman’s also been on record saying he can’t imagine raising a newborn without ChatGPT, though he did acknowledge people managed it for centuries. Barely, apparently.

Look, I get the appeal of outsourcing mental load. Parents are busy. Mornings are chaos. But there’s a difference between using AI to help you plan a birthday party and using it to generate a scripted audio experience about your own children so you don’t have to, you know, talk to them. The whole point of the drive to school is that it’s unstructured time. You hear about the kid who got teased at recess, the teacher who said something weird, the friend drama that’s actually a big deal to a nine-year-old. A podcast generated from a calendar isn’t going to capture any of that. It’s going to give you a sanitized, schedule-driven version of your family life, and you’ll listen to it instead of the real thing.

OpenAI is clearly trying to court parents. They’ve got a job listing up for a product manager focused on trust-sensitive consumer experiences for families. Meta’s testing an AI bedtime story app. There’s a market here, sure. But the company also faces multiple lawsuits from families who say ChatGPT contributed to delusions and suicides. They claim they’re working on making the models safer in sensitive interactions. That’s good. Maybe focus on that before you start narrating a child’s soccer game.

I’m not anti-AI. I use it for all sorts of things. But there’s a line between convenience and abdication. If you need a podcast to know what your kid is excited about, the problem isn’t a lack of AI. It’s a lack of asking. And no model, no matter how well-calibrated, is going to fix that.