I work with AI every day. Then I go home and try to help my 14-year-old plan a future I can't picture.

My stepdaughter is 14, which means we are entering the phase of parenting where everyone suddenly expects her to have opinions about her future. What classes does she want to take? What is she good at? Does she want to go to college? What kind of job might she want someday?

These are normal questions. We ask them because that is what you do with teenagers — you start helping them picture adulthood. The problem is that I am having an increasingly hard time picturing it for her.

Much of my professional life is spent researching and writing about AI, which has gone from an unsettling side conversation about the death of blogging and copywriting careers to something that touches nearly every aspect of our lives.

Then I go home and help raise a kid who will enter college and the workforce after all of these institutions have had several more years to rearrange themselves around AI. I am, in other words, probably more plugged into this stuff than is good for me.

I use AI every day at work. It's mandatory. I see what it can do well, what it still does terribly, and how quickly the gap between the two keeps changing. I read the research and follow the product releases. I watch businesses try to figure out what any of it means and watch my students figure out where the line is between using a tool and outsourcing their own brains. Despite all that exposure — or maybe because of it — I am anxious.

Last week, an Anthropic safety researcher said he puts the chance of AI killing everyone within the next decade at greater than 10%. If someone told me there was a greater than 10% chance a plane I was about to board would crash, I would not get on the plane. If someone tells me there is a greater than 10% chance AI kills humanity, I am supposed to read the article, close my laptop, and then proceed to ask a teenager whether she has thought about where she wants to go to college while making sure she eats enough vegetables and doesn't drink too much caffeine (add energy drinks to my list of ex

Maybe being a millennial is part of the problem. I am exactly the right age to have spent my formative years learning about the brand-new World Wide web while watching movies that explained that computers becoming too intelligent was generally a bad development. "The Matrix" came out when I was 12, "A.I. Artificial Intelligence" when I was 14, and "I, Robot" when I was 17.

We were not exactly given a lot of pop-cultural examples where humanity created increasingly intelligent machines, and everyone enjoyed a sensible increase in workplace productivity.

There is something genuinely surreal about growing up with artificial intelligence as one of our favorite dystopian plot devices, only to reach middle age just in time to hear actual AI researchers debating the probability of extinction.

Add BI in Google so our reporting is easier to find when you’re searching for what matters.

I am not convinced AI is going to kill us all. I am also not convinced the people warning about it are disinterested observers. Building a crisis only you can fix is textbook, and all PR is good PR. I also don't think I get to dismiss the warnings just because the incentives are messy.

AI systems are becoming more autonomous, more capable, and harder to predict. Researchers and policymakers are now openly debating risks that would have sounded like science fiction when my stepdaughter was born. Meanwhile, I am supposed to help her make choices based on an adult world that may look very different by the time she enters it.