No. 102 of 500 No. 14 in AI for Beginners

Artificial Intelligence

A Guide for Thinking Humans

By Melanie Mitchell

Published
October 15, 2019
Print length
336 pages
Reading level
Beginner
ISBN-13
978-0374257835
Publisher
Farrar, Straus and Giroux
Written for
General readers and professionals who want to understand what AI can and cannot do.

Synopsis

Melanie Mitchell, a computer scientist at the Santa Fe Institute, opens with her mentor Douglas Hofstadter's fear that advancing AI could leave human minds looking like relics, then sets out to judge how intelligent the leading systems really are. The book traces AI from its symbolic beginnings through the deep learning boom, explaining in plain language how convolutional networks classify images, how reinforcement learning mastered Atari games and Go, and how language systems handle translation and question answering. Along the way she talks with fellow researchers and documents where these systems break: adversarial images that fool classifiers, brittleness outside their training data, and bias inherited from datasets.

The closing section turns to the hardest problem: common sense, analogy and the background knowledge that lets people understand even a simple situation. Mitchell argues that the systems she examines lack these abilities, so claims of imminent superintelligence outrun the evidence; a shorter section addresses trustworthy and ethical AI. Published in 2019, the book predates ChatGPT, so its specifics on language models have aged, but its method of questioning benchmarks and probing failure modes applies directly to current debates. Readers finish able to separate demonstrated capability from marketing.

Scorecard

Editorial score

How it scored

Editorial score: 73.3 / 1005 weighted criteria

  • Influence30%

    22.5/30
  • Authority25%

    21.3/25
  • Practical value20%

    12.0/20
  • Currency15%

    9.0/15
  • Readability10%

    8.5/10
  • Editorial score

    73.3/100

How the rank was computed

Rank No. 102 of 500 comes from a composite of 63.0: observable signals 30.0 of 55 points plus editorial review 33.0 of 45 points.

All 11 inputs
InputPoints
Observable signals30.0/55
Reach (Amazon ratings volume)11.5/16.5
Consensus across best-of lists11.0/13.8
Awards and bestseller evidence0.0/8.3
Reader rating, sample-adjusted6.8/8.3
Amazon sales rank0.7/5.5
Recency0.0/2.8
Editorial review33.0/45
Author authority9.6/11.3
Influence10.1/13.5
Practical value5.4/9.0
Readability3.8/4.5
Currency4.1/6.8

Each criterion is scored from 1 to 10 and weighted, and the points add up to the editorial score. The rank combines observable signals (55 percent) with this score (45 percent). Read the methodology.

Verdict

A lucid, skeptical tour of how modern AI works and where it falls short, written by a respected complexity scientist.

Strengths

  • Clear explanations of neural networks for nonspecialists
  • Measured skepticism grounded in research experience
  • Strong chapters on analogy and common sense

Keep in mind

Because it was written before ChatGPT, its discussion of language systems is dated, and readers wanting policy or business guidance will find little here.

Key takeaways

  1. Benchmark success can mislead: image classifiers and game-playing systems that beat humans on narrow tests often fail when conditions shift slightly.
  2. Adversarial examples, tiny changes invisible to people, can make deep networks misclassify images with high confidence, exposing how differently they perceive the world.
  3. Common sense, abstraction and analogy, which humans use effortlessly, remain the central unsolved barrier between pattern recognition and general intelligence.
  4. Overestimating AI is itself a risk, because people may trust brittle systems in high-stakes settings before their limitations are understood.

Notable facts

  • Mitchell's website states the book was named one of the five best books on AI by both The New York Times and The Wall Street Journal. Source (opens in a new tab)
  • Reviewing it for Skeptic magazine, Peter Kassan called it the most intelligent book on the subject he had read; Publishers Weekly and The New Yorker also praised its accessibility. Source (opens in a new tab)

About the author

Melanie Mitchell

Melanie Mitchell is a professor at the Santa Fe Institute. She earned her PhD in computer science at the University of Michigan in 1990, studying under Douglas Hofstadter, and has held positions at Los Alamos National Laboratory and Portland State University. Her book Complexity: A Guided Tour won the 2010 Phi Beta Kappa Science Book Award.

Questions

Is Artificial Intelligence: A Guide for Thinking Humans still worth reading in 2026?
Yes, with a caveat about its date. The book was published in 2019, before ChatGPT, so its examples of language systems are dated. Its core arguments about brittle benchmarks, adversarial failures and the missing ingredient of common sense remain central to current debates about large language models, and Mitchell continues to develop them in her newsletter of the same name.
Do I need a technical background to read Artificial Intelligence: A Guide for Thinking Humans?
No. Mitchell explains neural networks, reinforcement learning and natural language processing without equations, using everyday analogies and concrete examples. Some passages on how networks learn require close attention, and Kirkus found a minority of the book abstruse, but most reviewers praised its lucidity. Readers with a technical background still gain from her research-informed perspective on where the field overreaches.
What is the main argument of Artificial Intelligence: A Guide for Thinking Humans?
Mitchell argues that people routinely overestimate AI. Systems that excel at narrow tasks such as image labeling, Go or answering quiz questions lack the common sense, abstraction and analogy-making on which human understanding depends. Until machines acquire those abilities, she contends, predictions of imminent superintelligence are premature, and the more pressing risk is trusting brittle systems too much.

Sources

As an Amazon Associate, Best AI Books earns from qualifying purchases. Best AI Books is sponsored by Iternal.ai, whose founder and CEO, John Byron Hanby IV, wrote The AI Strategy Blueprint, our #1 ranked book. How we rank books.