Book: Perspectives on Machine Consciousness

Perspectives on Machine Consciousness brings together 25 scientists and philosophers to ask whether artificial systems can have subjective experience, and what it would mean if they do. 

Perspectives on Machine Consciousness

Why this book, why now?

AI consciousness stopped being a niche philosophical question the moment frontier labs started taking it seriously. Anthropic created a dedicated model welfare role in 2024, and researcher Kyle Fish now leads work investigating whether Claude and similar systems might have morally relevant experiences. OpenAI's researchers and leadership have discussed the question in public interviews and essays, without landing on a settled position. Academic philosophers remain split: some, writing in journals like Nature's Humanities and Social Sciences Communications, argue there is no such thing as conscious AI even in principle. Others argue the opposite, or argue that the uncertainty itself demands caution.

Perspectives on Machine Consciousness puts 25 of these voices in one place: scientists, philosophers, and AI safety researchers who disagree with each other openly, including several who have not shared a platform before.

About the book

Perspectives on Machine Consciousness is a 26-chapter anthology edited by Ted Lappas and Calum Chace, published by Conscium, the AI safety company behind VerifyAX. The book opens with Daniel Hulme’s account of why intelligence and consciousness have been treated as one problem when they are two, and closes with Conscium’s own manifesto on why building AI capable of genuine experience may be safer than building AI that only optimises.

In between, contributors take positions that often contradict each other. Cameron Berg estimates a 25 to 35 percent probability that current frontier models have some form of conscious experience. Anil Seth argues the opposite. Susan Schneider identifies a “consciousness grey zone” that includes some hybrid computing architectures. The book does not resolve the disagreement. It puts the disagreement on the page.

A Selection of the Contributors

 

Contributor

Chapter

In brief

Daniel Hulme

Daniel Hulme

CEO of Conscium and founder of Satalia

The Spinning Wheel

Hulme argues that intelligence and consciousness are two separate problems. He proposes that consciousness emerges from integrated adaptive control, and sets out a Consciousness Risk Rubric and a Minimal Suffering Hypothesis for the people building AI.

Anil Seth

Anil Seth

Neuroscientist at the University of Sussex and author of Being You

A Sceptical View of Machine Consciousness

Seth makes the sceptical case. He ties consciousness to being a living body that works to keep itself alive, which puts machines out of reach whatever their intelligence.

Susan Schneider

Susan Schneider

Philosopher and author of Artificial You

The Quantum Darwinist Theory of Consciousness

Schneider sets out a quantum Darwinist theory of consciousness. She identifies a “consciousness grey zone” that takes in some hybrid computing architectures.

Mark Solms & Jonathan Shock

Mark Solms & Jonathan Shock

Neuropsychologist and author of The Hidden Spring, with physicist Jonathan Shock

How Could Consciousness Be Cultivated in a Machine?

Solms and Shock start from feeling. They set out how to build the simplest possible system that “gives a damn” about its own survival.

Karl Friston

Karl Friston

Neuroscientist at UCL and originator of the free energy principle

Artificial Consciousness and the Free Energy Principle

Friston applies the free energy principle to artificial consciousness. He defines a self through Markov blankets and self-evidencing, then asks what a machine would need to meet that definition.

David Brin

David Brin

Hugo and Nebula award-winning science fiction author

Selected, Blind Gropings of the Elephant of Consciousness

Brin argues that wanting, built by evolution, is the one thing AI still lacks, and the thing humans bring to any merger with it.

Cameron Berg

Cameron Berg

AI researcher studying consciousness in language models

The Evidence for AI Consciousness Today

Berg puts the probability that today's frontier models have some form of conscious experience at 25 to 35 percent. He bases the estimate on the models' internal dynamics rather than on what they say about themselves.

Calum Chace & Ted Lappas

Calum Chace & Ted Lappas

Editors

Hear From the Contributors

Daniel Hulme on why consciousness itself may be the missing safety mechanism for superintelligent systems

David Brin on Why Wanting Is the One Thing AI Still Can't Do

Cameron Berg on the evidence for AI consciousness today

Mark Solms on Why the Brainstem, Not the Cortex, Is Where Consciousness Begins

FAQs

What is Perspectives on Machine Consciousness about?

It's a 26-chapter anthology, edited by Ted Lappas and Calum Chace, in which scientists, philosophers, and AI researchers debate whether artificial systems have or could develop conscious experience, and what that would mean for AI safety and policy.

Are today's AI models conscious?

There's no consensus. Contributor Cameron Berg estimates a 25 to 35 percent probability that current frontier models have some form of conscious experience. Others, including Anil Seth, argue machines can probably never be conscious.

Why are AI companies like Anthropic and OpenAI talking about AI consciousness?

As AI systems have become more capable, frontier labs have started taking questions about machine welfare and consciousness more seriously. Anthropic created a dedicated model welfare research role in 2024. The topic remains unsettled across the industry.

Who contributed to the book?

25 scientists, philosophers, and AI researchers, including Anil Seth, Karl Friston, Susan Schneider, David Brin, and Conscium founder Daniel Hulme.

Where can I buy the book?

Perspectives on Machine Consciousness is available on Amazon UK.

 25 experts. No consensus. Read the arguments for yourself.