Reading list

Where to start

A short, opinionated shelf. Not everything ever written about machine minds — just the works we return to, in a rough order that makes sense if you are reading yourself into the subject. Each note says why it earns your time.

• Computing Machinery and Intelligence (1950)
  Alan Turing

  The essay that started it. Turing sets aside the unanswerable “can machines think?” and proposes the imitation game instead. Still the clearest thing ever written on why the question is so slippery — and remarkably, he anticipates and answers most of the objections that people are still raising today. Short, witty, and free online.

• Minds, Brains, and Programs (1980)
  John Searle

  The paper that introduced the Chinese Room. Whether or not you end up agreeing with Searle, you cannot think seriously about machine understanding without meeting his argument first. Read it alongside the published replies — the debate is more useful than the verdict.

• What Is It Like to Be a Bat? (1974)
  Thomas Nagel

  A dozen pages that reframed the study of consciousness. Nagel’s point — that there is a subjective “what it is like” which no third-person description captures — is the wall every theory of machine experience eventually runs into. Essential before you argue about whether AI could feel anything.

• Computer Power and Human Reason (1976)
  Joseph Weizenbaum

  The creator of ELIZA, alarmed by how readily people confided in his simple program, turns on his own field. A humane, prophetic argument about the difference between what we can compute and what we should hand to computers. The founding text for anyone worried about the god-shaped socket.

• The Human Use of Human Beings (1950)
  Norbert Wiener

  The father of cybernetics thinking about automation, control, and what machines would do to human society — decades early and often uncannily right. A reminder that the current debates are not new, and that the wisest early voices were the ones asking about ends, not just means.

• Gödel, Escher, Bach: An Eternal Golden Braid (1979)
  Douglas Hofstadter

  A sprawling, playful meditation on how meaning and “self” might arise from formal, meaningless rules — self-reference all the way up. Not a quick read, but nothing else conveys as vividly the intuition that mind could be a pattern rather than a substance.

• The Intentional Stance (1987)
  Daniel Dennett

  Dennett’s account of why we so naturally explain complex systems in terms of beliefs and desires — and what that habit does and doesn’t tell us about whether a mind is really there. The best antidote to sloppy talk about what an AI “wants.”

• The Conscious Mind (1996)
  David Chalmers

  The book that named the “hard problem” of consciousness and made the case that experience does not obviously reduce to physical function. Rigorous and readable; it maps the terrain any claim about machine sentience has to cross.

• Superintelligence: Paths, Dangers, Strategies (2014)
  Nick Bostrom

  The book that pushed long-term AI risk into the mainstream. Read it critically — some scenarios are more speculative than they sound — but it frames the control and alignment questions with unusual seriousness and rigour.

• Human Compatible: AI and the Problem of Control (2019)
  Stuart Russell

  A leading AI researcher’s calm, technical case that we have been defining the goal of AI wrong from the start, and how to fix it. More grounded and constructive than most books in the genre, written by someone who builds the systems.

• The Alignment Problem (2020)
  Brian Christian

  The best single narrative introduction to how modern machine learning actually goes wrong — bias, reward hacking, the gap between what we ask for and what we mean. Reported through the researchers doing the work, so the abstractions stay concrete.

• On the Dangers of Stochastic Parrots (2021)
  Bender, Gebru, McMillan-Major & Shmitchell

  The paper that gave us the phrase. A sharp, skeptical account of what large language models are and are not — a corrective to hype that is worth reading even, or especially, if you suspect the models are more than parrots.

• The Sovereignty of Good (1970)
  Iris Murdoch

  Not about AI at all — which is why it belongs here. Murdoch’s essays on attention, love, and moral vision are the clearest reminder of what the human questions were before machines got tangled in them. Read it to remember what all the arguing is ultimately about.