BOOK REVIEW
The Worlds I See: Curiosity, Exploration, and Discovery at the Dawn of AI, by Fei-Fei Li, Flatiron Books: A Moment of Lift Book, 02 September 2025, 336 pages, $18.48 (Paperback), ISBN 9781250898104
She is walking fast through a plain Washington hotel lobby, hearing her boots strike thin carpet like a metronome that has lost patience, trying to look calm while she feels anything but calm. In a few minutes she will sit at a witness table in the Rayburn House Office Building, her name printed in simple type, and testify about artificial intelligence, a topic that has suddenly become public property, fought over by lawmakers, companies, journalists, and activists. The tension in the opening is not the normal fear of public speaking, though she has that too, but the knowledge that her mother is in intensive care back in Palo Alto, that much of her testimony was drafted outside an ICU room, and that being an only child in an immigrant family means the world can ask for your expertise at the same time your private life asks for your body.
The Worlds I See: Curiosity, Exploration, and Discovery at the Dawn of AI is written by Dr. Fei-Fei Li, a Stanford professor and a leading researcher in computer vision and machine learning, who has also worked inside industry at a level that few academics ever reach. She served as Chief Scientist of AI and Machine Learning at Google Cloud while on sabbatical from Stanford, and she is a co-director of Stanford’s Institute for Human-Centered Artificial Intelligence, credentials that matter because this memoir depends on authority without sounding like a lecture. In the book, she also recounts co-founding AI4ALL, an education nonprofit meant to widen access to AI for students who are often shut out, which helps explain why this is not only a personal story but also an argument about who gets to build the future.
The memoir’s central idea is stated early and then tested across the rest of the book: it matters what motivates AI, and a human-centered motivation is not a slogan but a discipline. Structurally, the book combines life narrative with an accessible account of how modern AI emerged from older traditions, including the fits and starts, the periods of hype and disappointment, and the eventual convergence of data, algorithms, and computing power. Its scope is wide, but the method is intimate, with public milestones, like speaking before Congress, kept in tension with private crises, like translating for sick parents, working exhausting jobs, and learning to survive in English without losing her inner voice.
The first chapters return to China, and they do so not as nostalgia but as origin, because her earliest “worlds” are built from family temperament. Her father is portrayed as curious to the point of recklessness, a man who takes his child into parks and fields, chasing butterflies and stick insects, and who is late to her birth because he is birdwatching, then names her Fei-Fei from a word associated with flying. Her mother, by contrast, is an intellectual shaped by frustration and thwarted ambition, and the household becomes a place where curiosity is encouraged but also sharpened into urgency, as if learning is not just joy but escape.
Immigration is narrated as bodily disorientation and social demotion, beginning with arrival at JFK, the father missing for hours, the mother with twenty dollars, and the daughter discovering that classroom English does not help when life is moving around you at full speed. The family settles in Parsippany, New Jersey, where American space feels almost unreal, yet their apartment is cramped enough that her bed is wedged between the kitchen and dining area, a small fact that carries a whole sociology of downward mobility. What makes these scenes effective is that they refuse melodrama while still showing the exhaustion of constant translation, the price of long-distance calls, and the way scarcity turns even childhood into an accounting exercise.
Education becomes both refuge and battlefield, and the book is unusually clear about how immigrants experience school as a sensory storm. She describes brighter colors, heavier textbooks, loud bells, and social rules she cannot yet read, while at home she is pinned to dictionaries, translating every assignment word by word. A shocking episode of violence against an ESL student, triggered by a minor hallway contact, exposes a deeper fear that immigrants often carry silently: that safety is conditional, and that even the choice of which language to protest in can feel like a trap.
The memoir’s account of work, especially as a teenager, is one of its strongest contributions to the study of immigrants’ education because it shows how “grit” can be a euphemism for exhaustion. She takes underpaid, off-the-books jobs, including in a restaurant where she is scolded for running in the dining area, and she learns that immigrant communities can be supportive yet also harsh, treating imagination as waste because survival is so demanding. Yet she also shows the small economic wins that matter, like being able to deposit cash into a bank account and make phone calls home with a little less panic, a reminder that education is often held up by invisible labor.
A key figure enters in the form of Mr. Sabella, her math teacher, and the relationship is described with the kind of gratitude that does not feel staged. He challenges her, tutors her, shares books, and offers a safe room that becomes an unofficial “Math Lab,” giving her not just instruction but protection and recognition at a time when recognition can be lifesaving. When Princeton acceptance comes with significant financial aid, the moment reads less like a trophy than like a breach in her mother’s long history of being quietly told to stay in her place.
As the book turns toward science, it explains AI in a way that makes the field feel like a set of arguments rather than a single invention. Li traces older symbolic approaches and “expert systems,” the later shift toward machine learning, and the idea that neural networks learn patterns from examples rather than following explicit rules, emphasizing that progress depended on both conceptual breakthroughs and material conditions like data availability. These pages matter for readers in organizational studies and policy because they show how a technology is never just an idea, but also an ecosystem of funding, hardware, labor, and public appetite.
One of the most memorable scientific scenes happens in a dark lab with a cat’s visual cortex wired to electrodes, where neural activity becomes sound through loudspeakers, and perception feels suddenly physical, almost audible as a kind of weather. The experiment’s goal, reconstructing what the cat sees from intercepted signals, becomes a personal turning point because it convinces her that research is not merely schoolwork but a way of living inside questions. Importantly, even here the memoir ties scientific exhilaration to immigrant reality, as she takes calls from her mother about customers at the family dry-cleaning shop while she is doing lab work across the country.
A thematic analysis of the twelve chapters clarifies how the book keeps tightening its central tension, moving from the private problem of belonging to the public problem of responsibility. Chapter 1, “Pins and Needles in D.C.,” begins with anxiety before Congress and frames the moral question of what should motivate AI, while Chapter 2, “Something to Chase,” roots that question in childhood wonder and family displacement. Chapter 3, “A Narrowing Gulf,” centers immigration and the shrinking distance between her origins and the American research world she does not yet know is nearby, Chapter 4, “Discovering the Mind,” captures her turn from physics toward cognition and computation, and Chapter 5, “First Light,” uses evolutionary vision as a metaphor for how sensing creates competition and how perception reshapes life.
Chapters 6 through 8, “The North Star,” “A Hypothesis,” and “Experimentation,” read as the making of a scientist who is also an immigrant caretaker, always measuring distance between the lab and home. Across these chapters, her “North Star” is not only a scientific bet but also a moral orientation, as she learns that the best questions are rarely chosen in comfort and that experiments are shaped by what resources and data exist, not just by what is elegant. The transitions between chapters feel like shifts in scale, from personal aspiration to lab practice, from lab practice to social consequence, with each step making it harder to pretend that science can stay sealed off from the world.
Chapters 9 through 12, “What Lies Beyond Everything?,” “Deceptively Simple,” “No One’s to Control,” and “The Next North Star,” position her at the center of a field that has escaped the academy and entered daily life, including schools, hospitals, and markets. Her work with health care collaborators shows her commitment to AI that supports care rather than replacing it, and her conversations with nurses about surveillance anxiety highlight how even well-intended sensing can resemble workplace monitoring if institutions later weaponize it. The title “No One’s to Control” becomes the memoir’s hardest truth: AI is no longer “ours,” and the problem becomes governance, incentives, and dignity in environments where power rarely moves slowly.
Her most famous field-level contribution, ImageNet, is often described as a large, labeled image dataset and benchmark that helped make modern computer vision and deep learning take off, and accounts of its history commonly highlight ImageNet’s completion around 2009 and the ImageNet challenge results that accelerated the deep learning boom (Krizhevsky et al., 2012). This matters for immigrants’ education, women in STEM, and ethics of data annotation because it surfaces a basic fact that is often hidden in “AI progress” narratives: learning systems are trained on human-curated worlds, and those worlds reflect labor, judgment, and social bias. Her advocacy for widening participation in AI, including through AI4ALL, aligns with the memoir’s insistence that the future of AI depends on who gets to enter the room and who is asked to do invisible work outside it.
The memoir’s controversial edge is also its honesty: Li is optimistic about AI’s potential while being clear about harms like biased algorithms, surveillance, and the temptation to chase profit and speed instead of human benefit. Some readers will find friction in her movement between academia, public service, and a senior industry role, because it raises the question of whether human-centered ideals can survive inside corporate structures that reward scale and capture. Yet the book does not ask to be read as a corporate apology; it asks to be read as a record of how a scientist tries to keep faith with people while the world turns her research into a market.
Placed beside peer books on data power, The Worlds I See offers a different kind of authority, less prosecutorial and more confessional. Cathy O’Neil (2017) critiques of how large-scale, opaque algorithms can reinforce inequality, while Wiggins and Jones (2023) frames data as a historical force with technical, political, and ethical consequences, and Li’s memoir sits between them by showing the life behind the pipeline, from immigrant education to research culture to public debate. What her book adds is the claim that ethics is not an add-on to technical work but a lived condition, learned at bedsides, in classrooms, in low-wage jobs, and in rooms where you are asked to represent an entire field while your phone might ring with personal disaster.
Disclosure of interest
The author(s) confirm that there are no financial or non-financial competing interests.
Statement of funding
No funding was received.
References
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. Advances in neural information processing systems, 25.
O'neil, C. (2017). Weapons of math destruction: How big data increases inequality and threatens democracy. Crown.
Wiggins, C., & Jones, M. L. (2023). How data happened: A history from the age of reason to the age of algorithms. WW Norton & Company.
*****
Reviewed by:
Mayukh Mukhopadhyay
Executive Doctoral Scholar
Indian Institute of Management Indore
*****