Joanna Stern says AI voice tools were the keeper from her yearlong test

Tech reporter Joanna Stern spent a year pushing artificial intelligence into ordinary tasks, and the tool that stuck was not a household robot or a flashy gadget. She told CNBC Make It that AI voice tools became useful because they fit into moments when typing on a screen would be awkward, distracting or impossible.

For investors watching the AI buildout, Stern’s experience points to a practical adoption pattern: people may keep the products that save time inside routines they already have. In her case, that meant talking through ideas while driving, walking or moving between meetings.

Stern, a former Wall Street Journal reporter and founder of the tech media venture New Things, documented the experiment in her book, “I Am Not a Robot: My Year Using AI to Do (Almost) Everything.” CNBC Make It reported that she tested a wide range of AI products and experiences, including a lawn mower, sunglasses, household robots, a T-shirt folder and an AI boyfriend.

Some of those tools did not remain part of her routine, according to CNBC Make It. Voice-based AI did.

Why voice became the standout

Stern said she began using ChatGPT’s Voice Mode, a feature that lets users speak to the chatbot and hear responses back. “I was talking to ChatGPT in the car almost daily,” she told CNBC Make It, adding that she continued the habit after finishing the book.

While working on the book, Stern said she used ChatGPT and Claude to build what she called “BookBots,” customized AI assistants that could access her outlines, reporting materials and transcripts. She wrote that those bots helped with tasks including copyediting, number work and brainstorming.

She also used AirPods to speak with those tools while trying to keep her book work organized. Stern told CNBC Make It that speaking through ideas felt natural and could lead to another thought. She said the systems began to sound more human over the course of the year, while also feeling lower-pressure than speaking to another person because the chatbot was “not really judging me.”

In one example from her book, Stern described driving while full of ideas, with her son watching a movie in the back seat. She wrote that she talked through themes for the book with ChatGPT. Stern also acknowledged that the scene felt somewhat sad, since she was speaking to a bot while her child watched a screen, but said it helped her arrive at the beach with less anxiety about unfinished work.

How she still uses it

Stern told CNBC Make It that before the experiment she often listened to podcasts or music during commutes. Voice AI gave her a way to use that time to prepare instead, she said.

She said she still uses voice mode to shape interview questions and think through ideas while walking or driving in New York. The appeal, she told CNBC Make It, is focus: phone and computer screens bring notifications, while voice lets her talk through a single task.

Convenience also matters. Stern said a colleague may not be available late at night, and a five-minute gap between meetings may be too short for a real call. AI is available in those moments, she said, and can help her refine early ideas before bringing them to people.

The wider workplace context

Stern’s experiment comes as AI use at work becomes more common. The CNBC and SurveyMonkey Quarterly AI and Jobs Survey published in May found that 57% of employees use AI at work. About one-third of workers said AI made them more productive, and that share rose to 73% among people who use AI daily or weekly.

CNBC Make It also noted two cautions. AI systems can produce false information, often called hallucinations, so users should check their work. Psychologists have also warned that chatbots and AI companions should not replace human relationships.

Stern gave a similar warning about voice tools. “We need to remember they are not human,” she told CNBC Make It. “Let’s not all be fooled by that.”

This story draws on original reporting from CNBC.