Joanna Stern spent a year testing the AI future. Here’s what she learned.

In the introduction to her new book, I Am Not a Robot: My Year Using AI to Do (Almost) Everything, Joanna Stern describes a typical day in what could conceivably be all of our futures. Autonomous robots mow your lawn, fold your laundry, and drive you to your favorite restaurant. Algorithms monitor your vitals, write bedtime stories for your children, and whisper personalized positive affirmations as you fall asleep.

Or maybe not. One reason Stern decided to write this book was to discover just how revolutionary the AI revolution was shaping up to be. “Was my life really going to be better with AI everywhere,” she writes, “or was I about to find out that this whole AI thing was another bold Silicon Valley bet — less about progress and more about profit?” If for progress, better how and for whom? If for profit, well, we already know the answer to both of those questions.

I Am Not a Robot also glimpses into how the daily usage of various AI products will change not just the world, but ourselves. Like smartphones and social media, AI tools are destined to evolve from novelty into ubiquity, always at our fingertips and increasingly difficult to live without, whether we like it or not. But at what cost?

Fresh off finishing her book — and, acting on the advice from one of her chatbots, quitting her longtime job as a senior tech columnist at The Wall Street Journal — Stern spoke with Big Think to reflect on which (if any) aspects of the AI boom are worth the hype and whether the person who went into this self-experiment is different from the one who came out the other side.

Utopia or dystopia?

As the daughter of a multiple-time breast cancer survivor, who has a scary 39% chance of developing the disease herself, Stern was curious to learn how AI-powered applications like Koios DS and Transpara were being used in her yearly screenings. What she saw did not disappoint.

Trained on millions of mammograms, these applications can identify “patterns at the pixel level,” as Koios CEO Chad McClennan put it. These are patterns the human eye might overlook, especially if a doctor is coming down with the flu or distracted by drama at home. Citing research published in the Journal of the National Cancer Institute, Stern writes that early detection through AI tools could reduce certain cancer risks by up to a third.

She was less enthusiastic after visiting the dentist, where AI-powered dental tools are being used to detect the early signs of cavities and more. Unlike Koios and Transpara, though, these tools actually made her feel less at ease, not more. Dentistry is not as regulated as cancer treatment, she explains, and many dentists told her they felt pressured by their employers to use AI to over-diagnose.

Dentistry reveals an important drawback to AI use in healthcare. “Since humans built these healthcare systems in the first place, with all their human incentives baked in,” Stern writes, “shouldn’t we fix the systems themselves before unleashing machines that will only magnify the issues?”

Even if the AI boom proves to be a bubble — one that bursts or merely deflates — Stern feels confident certain uses of the technology can no longer be put back in the bottle. “Generative AI to search for and obtain information is certainly here to stay,” she tells Big Think. “People have come to expect quick answers versus a huge list of links, and there’s no going back.”

To this list, she would also add self-driving cars and AI assistants. Despite her Waymo stopping in the middle of a highway because it thought the person in the car ahead was outside the vehicle rather than inside, or describing her attempt to let AI compose her emails as the “fast track to unemployment and divorce,” we seem close to approaching a point where their general usefulness will outweigh their liabilities.

But while Stern’s experiment left her feeling hopeful, that doesn’t mean we should buy everything Big Tech tries to sell us. “A lot of what’s possible now wasn’t possible even two years ago,” she says. “Still, we should be skeptical of anyone who speaks about the future of AI with total certainty. Progress is real, but also unpredictable — even for the executives making the biggest decisions.”

Assholes and AI elbows

In the days of Nokias and Blackberries, mobile phones were largely limited to calls, texts, and the occasional game of Snake. Nowadays, of course, we use our smartphones for nearly everything: shopping, directions, watching Netflix, booking hotels, playing games, monitoring our homes, and on and on.

As these devices play ever larger roles in our lives, so too does their impact on how we think, feel, and relate to each other. Various studies show that smartphone and social media use negatively impact our cognitive abilities, attention spans, and independent, sustained emotional regulation, among other things.

AI, for its part, might be turning us into jerks. By now, everyone has cursed at an AI chatbot. If you haven’t, then you’ve definitely wanted to. Stern has, and she’s not sure how to feel about it. “Would I berate a human this way?” she asked herself after tearing ChatGPT a new one for refusing to understand (let alone admit) that the image she had asked it to generate for one of her bedtime stories shows six hamsters instead of five.

“The maddening part is not just that it’s wrong,” she tells Big Think, “but that it’s so confident that it’s right. There’s something infuriating about being gaslit by a machine that has access to so much of the world’s knowledge.”

However maddening AI can be, we shouldn’t be too quick to take our anger out on it. Not because we’ll hurt its feelings — it has no feelings to hurt — but because it might lead us to behave similarly to actual people. As Jennifer Siegel, a research associate at Bentley University’s User Experience Center, notes in an article on the subject:

“We may argue that abusing AI is a way of letting off steam on an inanimate object, rather than an actual person. However, studies have shown that venting does not reduce emotion, but rehearses it (…) Therefore, practicing this behavior towards virtual assistants could bleed into our interactions with other beings.”

One way Stern has learned to keep her AI rage in check is by changing how she conceives of her AI tools. “Instead of expecting AI to be an expert, I started treating it like an eager intern. Sometimes it produces work that just blows me away. Other times it needs careful supervision, fact-checking, and a break.”

Cognitive offloading

AI may also make us lazier and dumber, much like smartphones and personal computers, only to a much greater degree. The technology is creating unprecedented opportunities for cognitive offloading: the delegation of mental labor to external resources.

Cognitive offloading certainly has its perks. Resources such as Google Maps and Wikipedia have broadened our horizons (literally and figuratively, respectively) and let us do things we couldn’t do before. However, it also comes with serious downsides. The more we rely on Google Maps, the less we come to know our hometowns. The more we rely on ChatGPT or Claude, the less we practice the skills that they are meant to augment.

This fear that technological innovation inevitably comes at the cost of human ingenuity is far from new. In ancient Greece, for example, people worried that writing things down would worsen their working memory. Similarly, painters feared that the advent of photography would end their art form.

And, in a way, they were right. Writing allowed people to record facts and ideas onto scrolls for later retrieval. While cameras didn’t spell the end of painting, they did change the art’s place in society, as well as pursuits like zoology and botany. The much broader applications of AI have the potential to automate all manner of activities, from writing and coding to patient care and customer support.

Of course, just because AI is capable of replacing these jobs doesn’t mean it should. Not only because replacement will put a lot of people out of work, but also because, in many cases, AI would be of more use to us when it complements human expertise rather than replaces it.

This is especially obvious in healthcare. Stern cites a study that found endoscopists’ ability to spot precancerous growths during colonoscopies without AI assistance dropped by more than 20% in the months after they started using these tools in their practices. In this case, the full benefits of AI are realized only if doctors use it to provide a second opinion, not as a substitute for their own judgment.

After her AI year, Stern is no stranger to the dangers of cognitive offloading. “I am far more dependent on AI than I was before,” she admits. “There are certain tasks, [such as] research and editing simple copy, where my muscle memory is simply to tap the app. If you took those tools away from me, I’d be annoyed. Really annoyed.”

“I’m not sure I’ve found the perfect solution,” she continues, “but I have started implementing ‘no AI’ tasks and times. There are things I want to force myself to do without AI: answering interview questions like these, writing my newsletter, thinking through the big strategy or message for a project. I can feel myself getting lazy, and I have to remind myself that if I don’t do that work myself, the end result won’t be as creative or as me.”

This article is featured on Big Think.