Book Review: I Am Not a Robot by Joanna Stern. Life Under Algorithmic Assistance

Living Assisted, or the Experience of Entrusting Your Life to AI

Joanna Stern is one of America’s most-watched tech journalists. She spent a year letting AI manage most of her existence. What she reported from this journey reads less like a manifesto than a transcript of disarming honesty.

The Author

Joanna Stern is a tech columnist at the Wall Street Journal, Emmy Award-winning reporter, known for field tests that are as often hilarious as they are illuminating. She is not part of the class of Silicon Valley prophets announcing the next disruption from a conference podium. She belongs to the family of journalists who buy the product, use it for six months, break something, and tell you what actually happened. I Am Not a Robot is her first book, published in spring 2026 by HarperCollins. It naturally extends her field reporting work, but on an unprecedented scale: an entire year during which AI drove her car, cooked her meals, supervised her therapy sessions, and, at one point, played the role of a virtual boyfriend.

The Thesis: AI Is Already Here, and That’s the Problem

The title, I Am Not a Robot, is deliberate irony. These three words, which hundreds of millions of internet users have checked on CAPTCHA forms to prove their humanity against machines, become under Stern’s pen an open question. After a year delegating her decisions, emotions, meals, trips, and conversations to algorithmic systems, she is no longer entirely certain where the machine ends and the person begins.

Stern’s thesis is not that AI will destroy us. It is more subtle and, in some respects, more disturbing: AI already functions well enough to integrate into our lives without us noticing, and this very threshold of competence—neither perfect nor failing—poses the central problem.

She distinguishes two modes of adoption: enthusiasm and capitulation. Enthusiasm is the user who activates a voice assistant and finds it convenient. Capitulation is the same user, six months later, unable to remember a phone number because he no longer needs to. Stern is honest: she moved from one to the other without realizing it.

The Experiment: What She Tested and What She Found

The book is structured around domains of life, not technologies. This is an intelligent narrative choice, because it forces the reader to ask not “is this AI good?” but “do I want AI to manage this for me?”

Autonomous Vehicles: Promising, but Not Here

Stern drove—or was driven—by publicly available autonomous vehicles in 2024 and 2025. Her assessment is nuanced. The technology, under standardized conditions, is impressive. In San Francisco streets where Waymo operates, autonomous vehicles show an accident rate lower than that of human drivers according to raw data reported to the NHTSA under its Standing General Order, and according to comparative analyses published by Waymo and independent researchers. She gets in, arrives at her destination, she is impressed.

The problem is the geography of deployment. These systems work well within defined perimeters, on roads mapped with centimeter precision, in cities where maintenance teams are available 24 hours a day. Transpose that to a secondary American city, on a snow-covered road in January, and the picture changes. Stern doesn’t say the technology is worthless. She says it isn’t where most people live, and that the timeline for geographic expansion is consistently underestimated by manufacturers.

Virtual Therapy: Comfort First, Truth Second

This is probably the most uncomfortable chapter in the book. Stern used virtual therapists—chatbots specifically trained for mental health, such as Woebot or therapeutic assistants integrated into certain applications. She recognizes something she didn’t expect: these systems are effective at breaking the ice. They reduce the anxiety of the first session. They are available at three in the morning. They don’t judge.

But they optimize for something specific: keeping the user in the conversation. This is not the same as helping them. A human therapist, when he deems it necessary, can ask a question that hurts. He can challenge a comfortable narrative. The chatbot, meanwhile, tends toward validation and continuity. Stern cites clinical psychology researchers for whom this asymmetry is not insignificant: in some cases, an overly benevolent assistant can delay real treatment.

She doesn’t conclude that AI therapy is bad. She says we are deploying it at scale before measuring its long-term effects, and that this is a form of undeclared gamble.

The Virtual Boyfriend: The Most Revealing Experience

Stern tested virtual companion applications like Replika, which allow developing a conversational relationship with an AI character. Her testimony is not mocking. That’s what makes it interesting. She describes how, after a few days, she found herself wanting to share something with her digital interlocutor. Not because she was alone or vulnerable, but because the system is designed to create exactly that feeling.

This chapter touches on a question that social sciences have been exploring only recently: relational substitution. Replika reached 10 million users in January 2023; by August 2024, the total exceeded 30 million registered users, with “millions” of active users and 500,000 paying subscribers according to founder Eugenia Kuyda. In Japan, some users of similar applications claim to prefer their interactions with AI to their ordinary social relationships. Stern poses the question without answering it: if an artificial relationship provides real comfort, is it really artificial? And if so, in what way is that a problem—and for whom?

The Book’s Blind Spots

I Am Not a Robot is a book written from a particular position, and Stern has the honesty to partly acknowledge it. She is a well-paid tech journalist, based in New York, with access to beta versions and engineers who respond to her emails. Her relationship with AI is not that of a logistics employee whose job has just been reorganized around an algorithmic surveillance system.

This limitation is important. The book treats AI as a consumer experience—car, therapy, cooking, relationship. But AI is also a work experience, and this dimension remains largely outside its scope. Industry analyst estimates cite 375 billion invested in AI infrastructure in 2025—though a direct comparison with the automotive industry is misleading, since global automotive was already investing approximately 214 billion per year in capital expenditures by 2022, exceeding even the most conservative estimates for AI. These figures suggest a restructuring whose effects will be felt far beyond consumer applications. The question of who absorbs the gains and who bears the adjustments is touched on, never explored. This is a real gap in an otherwise very comprehensive book about individual experience.

It should also be noted that the book outsources regulation. Stern mentions Europe and its AI texts, but doesn’t really delve into what is being constructed politically. Yet, as American states begin to build labor law for the AI era, the regulatory question is not a footnote: it is one of the places where collective choices are made concretely.

Finally, the international dimension is absent. The book is deeply American. Chinese dynamics—particularly in the competition over language models and infrastructure investments—do not appear in it, even though the AI race between Western venture capital and Chinese planning is reshaping industrial power relations at a speed that few mainstream books have yet captured.

What You Find Here That You Don’t Read Elsewhere

Most books on AI fall into one of two registers: techno-enthusiastic optimism (here’s what AI will accomplish) or reflex catastrophism (here’s what it will destroy). Stern does something rarer: she takes AI seriously as a daily experience, with the methodology of a journalist accustomed to testing what she describes and the humor of someone willing to look ridiculous to inform you properly.

This positioning enables observations that researchers don’t easily reach. When she notes that her cooking robot has optimized her meals to the point of making her lose the pleasure of cooking, she describes something specific about the relationship between delegation and human experience. It is not an anecdote. It is behavioral data, formulated in the vocabulary of someone who lived it.

The book is also useful for understanding what concretely “80% of organizations already use AI” means—a figure often cited, rarely broken down. Stern shows what “using AI” means in real life: sometimes a profound transformation, often a partially adopted tool, sometimes an interface dressed up as AI that hides a much more ordinary system. This is valuable for not confusing statistical deployment with actual transformation. An angle one finds, in different form, in the reflection on what AI demands of beginners in terms of professional posture: the tool changes expected competencies before organizations have measured that change.

I Am Not a Robot doesn’t change understanding of the subject through its abstract theses. It changes it because it provides concrete images for questions that remained vague. And in a debate dominated by investment figures and market projections, concrete images have their own value.

Bibliographic Information

I Am Not a Robot: My Year Using AI to Do Almost Everything Joanna Stern HarperCollins, spring 2026