Best Philosophy of AI Books: Where to Start, Ranked

Every few months a new chatbot reopens questions that philosophers have argued about since Alan Turing’s 1950 paper “Computing Machinery and Intelligence.” Can a machine think? Can it understand what it says? Could it feel anything, and if so, what would we owe it? And what happens to us when we hand our judgment, our writing and our attention to systems we don’t understand?

Our philosophy of technology ranking already covers the big books on AI as a social and political force, including Shannon Vallor’s The AI Mirror, Matteo Pasquinelli’s The Eye of the Master and Kate Crawford’s Atlas of AI. This list asks the older and stranger question about minds. We ranked books by philosophical depth, clarity, and how well they’ve aged as the technology changed under them.

Where to start

Start with Brian Cantwell Smith’s The Promise of Artificial Intelligence and Hubert Dreyfus’s What Computers Still Can’t Do for the deep question of understanding. Read Jonathan Birch’s The Edge of Sentience for the ethics of uncertainty, and Bostrom and Russell for the risks.

1. Brian Cantwell Smith, The Promise of Artificial Intelligence: Reckoning and Judgment (2019)

The deepest short book on what AI can and can’t do. The Promise of Artificial Intelligence distinguishes “reckoning,” the calculative power at which machines now excel, from “judgment,” a committed, responsible engagement with the world as it actually is. Smith, a computer scientist turned philosopher, doesn’t say machines can never have judgment. He says we should stop mistaking one for the other. Every AI debate would improve if both sides read it.

2. Hubert L. Dreyfus, What Computers Still Can’t Do (1972; revised 1992)

The classic critique, and still the best-argued. Drawing on Heidegger and Merleau-Ponty, Dreyfus argued that human intelligence rests on embodied know-how and a background of practical understanding that can’t be written down as rules. What Computers Still Can’t Do was aimed at the symbolic AI of its day, and the turn to machine learning vindicated part of it. The question it leaves us is whether statistical learning from human data gets around the problem or only hides it.

3. Jonathan Birch, The Edge of Sentience: Risk and Precaution in Humans, Other Animals, and AI (2024)

The most careful book on the moral question. The Edge of Sentience starts from the fact that we often can’t know whether a being is sentient, whether that being is a patient with a brain injury, an octopus or a future AI system, and asks how we should act under that uncertainty. Birch builds a framework of precaution rather than proof. It’s rigorous, humane and directly relevant to policy.

4. Nick Bostrom, Superintelligence: Paths, Dangers, Strategies (2014)

The book that turned “AI risk” into a mainstream concern. Superintelligence argues that a system much smarter than us could pursue almost any goal, and that controlling it would be extraordinarily hard. You don’t have to accept its timelines or its thought experiments to see why it shaped a decade of argument. Read it with the critics beside it.

5. Stuart Russell, Human Compatible: Artificial Intelligence and the Problem of Control (2019)

A leading AI researcher’s answer to Bostrom. Human Compatible proposes that machines should be built to be uncertain about what humans want, so that they defer, ask and allow themselves to be switched off. It’s the clearest case that the control problem is an engineering and philosophical problem rather than science fiction.

6. Christopher Summerfield, These Strange New Minds: How AI Learned to Talk and What It Means (2025)

The best recent book on large language models. Summerfield, an Oxford cognitive neuroscientist who has also worked in AI research, explains in These Strange New Minds how chatbots learned to talk, and he takes seriously the question of whether “understanding” is the right word for what they do. It’s balanced, current and refreshingly free of hype in either direction.

7. N. Katherine Hayles, Bacteria to AI: Human Futures with Our Nonhuman Symbionts (2025)

The most ambitious new theory on the list. In Bacteria to AI, Hayles, author of the classic How We Became Posthuman, proposes an “integrated cognitive framework” in which meaning-making extends from bacteria and plants through animals and humans to some AI systems. It’s demanding, and it reframes the whole debate: the question becomes what kinds of cognition we live among, not whether machines can think “like us.”

8. Meghan O’Gieblyn, God, Human, Animal, Machine (2021)

The best literary book on AI. A former student of theology, O’Gieblyn shows in God, Human, Animal, Machine how talk about technology recycles religious ideas about the soul, resurrection and transcendence. It’s beautifully written and quietly devastating about transhumanism.

9. Susan Schneider, Artificial You: AI and the Future of Your Mind (2019)

A philosopher of mind asks what happens if we merge with machines. Artificial You proposes tests for machine consciousness and asks whether uploading or brain enhancement would preserve the self or quietly end it. It’s short, lively and a good introduction to the personal-identity side of the debate.

10. Margaret A. Boden, AI: Its Nature and Future (2016)

The best compact overview, by a pioneer of cognitive science. AI: Its Nature and Future explains the main approaches to AI and the philosophical questions each raises. It predates the chatbot boom, but it gives the conceptual map that makes the newer books easier to follow.

Honorable mentions

Hegel in a Wired Brain by Slavoj Žižek (2020): what happens to subjectivity when the brain is wired directly to machines.
The AI Mirror by Shannon Vallor (2024): AI as a mirror that reflects our past back at us.
Reality+ by David Chalmers (2022): virtual worlds and simulated minds, defended as genuinely real.

The continental view: Žižek, Han, Agamben and Sloterdijk

Analytic philosophers ask whether machines can think. Continental philosophers tend to ask what thinking machines do to us. Slavoj Žižek’s Hegel in a Wired Brain asks whether a mind directly linked to a machine would still be a subject at all, since for Žižek the subject is a gap and not a store of information. Byung-Chul Han’s Non-things argues that artificial intelligence can’t think, because thinking begins with being emotionally gripped by the world, and AI can’t get goosebumps. Giorgio Agamben’s What Is an Apparatus? supplies the theory of devices that capture and shape living beings. Peter Sloterdijk’s You Must Change Your Life reminds us that humans have always been shaped by techniques of practice and training, and his “Rules for the Human Zoo,” collected in Not Saved, asks who gets to do the shaping.

One to watch: Rafael Gallardo

Most books on AI ask whether machines will come to have desires of their own. Rafael Gallardo, writing at The Rhizome Times, worries about the reverse: that our desires are becoming theirs. In “Unknown Known,” he argues that platforms capture our “main activity,” so that “We are left living out, through our activities, the desires of others, now the desires of artificial systems.” He calls this condition “the reality of the virtual”: information and algorithms forming an invisible architecture that shapes perception, desire and identity. His vocabulary comes from Žižek’s reading of Lacan and from Mark Fisher, and it brings a psychoanalytic question to a debate dominated by engineers. You can now preorder his first book, Indoctrination Guide, in a digital edition due December 25, 2026. Read him at The Rhizome Times.

Worth subscribing

L. M. Sacasas, The Convivial Society: essays on technology, culture and the moral life, grounded in the history and philosophy of technology. Start with “AI Is Not Conscious, But It Is Becoming Our Unconscious”. 48K+ subscribers on Substack.
Erik Hoel, The Intrinsic Perspective: a neuroscientist and consciousness researcher turned essayist, writing on AI, consciousness and culture; see “Don’t dethrone consciousness!”. His book The World Behind the World (2023) is on consciousness and free will. 72K+ subscribers on Substack.

Conclusion

Read Smith and Dreyfus for the deep question of understanding, Birch for the ethics of uncertainty, and Bostrom and Russell for the risks. Then read Summerfield and Hayles to see how the debate looks now that machines really do talk back. The question “can machines think?” turns out to be inseparable from a harder one: what were we doing when we called it thinking?

A quick note on links: Philosophy Ranked is reader-supported. When you buy a book or subscribe to a writer through links on this page, we may earn a small commission or referral fee, at no extra cost to you. As an Amazon Associate I earn from qualifying purchases.