Home / Courses / Consciousness, AI, and the Future of Mind

🤖

Consciousness, AI, and the Future of Mind

Can machines be conscious? And what does that mean for us?

7 modules 14 hours Intermediate
Explores the intersection of artificial intelligence and consciousness studies. Topics include the Chinese Room argument, the orthogonality thesis, phenomenal vs. functional consciousness in LLMs, the alignment problem, ethics of synthetic consciousness, and what neuroscience tells us about creating artificial general intelligence.

What you'll learn

  • Evaluate the Chinese Room argument and its implications for AI consciousness
  • Identify criteria for determining whether an AI system could be conscious
  • Navigate the debate over large language models and phenomenal consciousness
  • Assess quantum and computational approaches to consciousness in AI
  • Understand superintelligence risks, the alignment problem, and AI ethics
  • Apply insights from disorders of consciousness to the challenge of detecting AI awareness
  • Develop your own informed position on machine consciousness and its implications

Course modules

Module 1

The Chinese Room — The Founding Argument

John Searle’s Chinese Room argument (1980) is the most famous — and most contested — argument against the possibility of genuine AI consciousness. By imagining himself inside a room manipulating symbols without understanding them, Searle claimed to show that syntax (symbol manipulation) is not sufficient for semantics (meaning, understanding, consciousness). This module examines the argument, its premises, the major responses, and why it remains central to the AI consciousness debate.
Required
Optional

Key thinkers in this module

John Searle Chinese Room; biological naturalism

🧠 Reflect: Does the Chinese Room argument refute the possibility of strong AI, or does it merely show that Searle — a human — doesn't understand Chinese? Could a different architecture overcome the objection?

Next → What Would It Take for an AI to Be Conscious?
Module 2

What Would It Take for an AI to Be Conscious?

If the Chinese Room doesn’t settle the question, what would? This module examines proposed criteria for AI consciousness: from the AI Consciousness Test (Schneider) to Chalmers’ framework for assessing phenomenal consciousness in artificial systems. We ask whether consciousness requires biological substrates, whether functional equivalence is sufficient, and what we should expect from a genuinely conscious AI. No position is asserted — competing views are presented for evaluation.
Required
Optional

Key thinkers in this module

David Chalmers Hard Problem; zombie argument; naturalistic dualism; panpsychism

🧠 Reflect: Imagine we build an AI that passes every behavioural test for consciousness but lacks a biological brain. Is it conscious? What further evidence could change your mind?

← Previous The Chinese Room — The Founding Argument Next → Large Language Models — The Current Debate
Module 3

Large Language Models — The Current Debate

The rapid advance of large language models (GPT-4, Claude, Gemini, and others) has brought the AI consciousness question from philosophy seminars to headlines. Proponents argue that LLMs exhibit reasoning, creativity, and self-awareness. Sceptics counter that they are stochastic parrots — sophisticated pattern-matchers without genuine understanding. This module presents the competing positions with the care they deserve, drawing on interviews and analyses from leading researchers across both camps.
Required
Optional

Key thinkers in this module

David Chalmers Hard Problem; zombie argument; naturalistic dualism; panpsychism Adrian Owen fMRI covert awareness in VS; CMD detection

🧠 Reflect: When an LLM says 'I am conscious', is that evidence of consciousness — or merely a statistical prediction of what a conscious being would say? How could we distinguish the two?

← Previous What Would It Take for an AI to Be Conscious? Next → Computation, Quantum Physics, and Consciousness
Module 4

Computation, Quantum Physics, and Consciousness

Roger Penrose argues that consciousness is not computational — that there is something in the nature of conscious understanding that cannot be captured by any algorithm. Drawing on Gödel’s incompleteness theorems and quantum physics, Penrose proposes that consciousness requires non-computable quantum processes in the brain’s microtubules (Orch-OR). This module critically examines Penrose’s arguments and their implications for the possibility of AI consciousness.
Required
Optional

Key thinkers in this module

Roger Penrose Non-computational consciousness; Orch-OR

🧠 Reflect: If Penrose is right that consciousness requires non-computable quantum processes, then classical AI (including LLMs) cannot be conscious. But is he right? What would settle the question?

← Previous Large Language Models — The Current Debate Next → Superintelligence — Risks, Alignment, and Ethics
Module 5

Superintelligence — Risks, Alignment, and Ethics

Whether or not AI is conscious, advanced AI systems will pose profound challenges: the alignment problem (how to ensure AI goals align with human values), the orthogonality thesis (intelligence and goals are independent), and the existential risks of uncontrolled superintelligence. If advanced AI were also conscious, the ethical stakes would be even higher — raising questions about AI rights, suffering, and moral status. This module maps the landscape of AI risk and ethics.
Required
Optional

Key thinkers in this module

Anil Seth "Controlled hallucination"; predictive self David Chalmers Hard Problem; zombie argument; naturalistic dualism; panpsychism

🧠 Reflect: If we create a superintelligent AI that is not conscious, does it have moral status? If it is conscious, does it have rights? Can one exist without the other?

← Previous Computation, Quantum Physics, and Consciousness Next → Detecting Consciousness — From Disorders of Consciousness to AI
Module 6

Detecting Consciousness — From Disorders of Consciousness to AI

How do we detect consciousness in beings that cannot communicate? This question is central both to clinical neurology (patients with disorders of consciousness) and to AI consciousness. This module examines methods developed for detecting awareness in vegetative and minimally conscious states — including fMRI, EEG, and the Perturbational Complexity Index (PCI) — and asks whether these techniques can be adapted as ‘consciousness tests’ for artificial systems.
Required
Optional

Key thinkers in this module

Adrian Owen fMRI covert awareness in VS; CMD detection Steven Laureys VS/MCS differentiation; UWS renaming Giulio Tononi IIT; Φ metric; PCI clinical tool

🧠 Reflect: If we adapted the PCI for AI systems, would a high complexity score be evidence of consciousness? Or could a non-conscious system also produce complex information integration?

← Previous Superintelligence — Risks, Alignment, and Ethics Next → The Future of Mind — Humans, AI, and Beyond
Module 7

The Future of Mind — Humans, AI, and Beyond

The final module looks ahead. What kind of future are we building? Will AI amplify or diminish human consciousness? Could mind uploading preserve personal identity? Should we be accelerating or slowing the development of artificial consciousness? Drawing on philosophy, cognitive science, and ethics, this module invites you to articulate your own vision for the future of mind — one that integrates what we’ve learned about consciousness with the transformative potential of AI.
Required
Optional

Key thinkers in this module

Philip Goff Panpsychism for general audience; Russellian monism Mihaly Csikszentmihalyi Flow theory; optimal experience Anil Seth "Controlled hallucination"; predictive self

🧠 Reflect: After seven modules exploring AI and consciousness, what is your position? Should we be trying to create conscious AI, or should we focus on making AI powerful but not conscious? What is at stake in this choice?

← Previous Detecting Consciousness — From Disorders of Consciousness to AI

📖 Study independently: All readings link to library entries on this platform. Full enrolments with video, quizzes, and certificates will be added in a future phase. View the roadmap →