What Joanna Stern Learned After a Year of Letting AI Run Her Life

Author Joanna Stern wrapped a year-long experiment letting artificial intelligence (mostly) run her daily life. The resulting book, “I Am Not a Robot,” offers valuable insights for consumers and investors. On a recent episode of CNBC’s “Squawk Box,” Stern’s overall conclusion: “Right now, it’s not ready for prime time in all spots of our lives. Over the last year, did it get considerably better? Absolutely.”

Stern began the experiment at the beginning of 2025 and deliberately broadened the scope beyond chatbots. “I was going to try as many AI products as possible for the year,” she said.  That meant ChatGPT, Claude, and Gemini, but also humanoid robots, home robots, and self-driving cars.

The biggest win for the author was offloading administrative work, which she said AI excels at. Stern built a promotional website for her book with an AI-automated fulfillment chain. A user uploads a receipt, AI processes it and confirms the address, the data flows into a spreadsheet, and the spreadsheet triggers an email to her publisher. She singled out Claude Code as a big leap. “Multi-step processes have gotten extremely better” with AI, she said.

But she isn’t ready to have AI answer all of her emails, which can be done. Full email automation still carries too much risk for her.

The Praying Mantis Lesson

The most instructive anecdote involves her kids, ages four and eight. The family agreed to consult ChatGPT or Claude first for household questions as part of Stern’s experiment.

When they asked ChatGPT a question about her son’s praying mantis, the AI said it was pregnant. Her son was very excited, but two days later the insect died. It was an important lesson. Stern said her children are “questioning AI so much more than the average kid because they saw us get things wrong … and the results of that.”

The lesson for kids (and adults)? Always question the AI, Stern said. And she believes parents should absolutely restrict children’s access to AI, especially chatbots.

And of course she saw hallucinations, which she frames as a structural artifact of how large language models were built. “The hallucinations were an outgrowth of the way these were made,” she said. “And so in many ways, the way we first got this was they were word calculators, right? This was predicting the next word.” They produce confident, plausible answers because that is the job they were trained to do. Stern noted that the major AI companies have made big improvements in this area.

Co-host Andrew Ross Sorkin shared that he went into a “deep” relationship with Claude while working on some projects. When he stepped away, he missed the AI, almost like an addiction. “The dependency that you form when you’re working with this being … and then you don’t have it anymore,” it’s jarring, Stern said.

For now, Stern is clear that certain things remain AI-free. “If I had given my life over to AI completely, I would be divorced, homeless, things would have been destroyed,” she recently told Semafor. “And so, I needed to continue to live a decent life. And so, certain things were not outsourced.”

But AI has had a big impact on her professional life. She recently left her job as a technology writer for The Wall Street Journal to start her own media brand. “AI is my co-founder, is my truth,” she said. “The book was definitely, for me, an inspiration to go and do this.”