Consciousness, Plants and AI: Anil Seth in Conversation with Michael Pollan at the Royal Institution

Overview

In this Royal Institution event, cognitive scientist Anil Seth interviews author Michael Pollan about consciousness, its definitions, and the boundaries between life, sentience, and machine intelligence.

• Consciousness is discussed as subjective experience, echoing Nagel's what it is like to be a bat.
• Pollan and Seth consider whether plants show sentience and how that relates to consciousness.
• The conversation moves to embodiment and how feelings, body states, and the brain shape experience.
• AI consciousness and the difficulty of separating software from hardware are debated, including thought experiments and ethical implications.

Audience questions from theatre and livestream attendees round out the discussion.

Introduction and framing

The dialogue opens at the Royal Institution with Anil Seth introducing Michael Pollan and outlining the plan of a live conversation that will later invite audience questions. The central goal is to explore what consciousness refers to, how we define it, and what the discussion implies for science, ethics, and technology.

Defining consciousness

Seth emphasizes subjective experience as the core of consciousness, drawing on Thomas Nagel’s famous question, what is it like to be a bat, to illustrate how consciousness may require a first person perspective rather than solely objective behavior. Pollan adds nuance by acknowledging the varied forms of life and the difficulty of drawing a sharp boundary between conscious and non conscious systems.

From plants to sentience

The discussion takes an unconventional turn by starting with plants. Pollan explains plant neurobiology experiments that show learning, memory, signaling, and interplant communication, though Seth cautions against equating these with human interiority or inner experiences. The conversation distinguishes sentience (environmental awareness and action) from consciousness (subjective experience), proposing a spectrum of life forms that solve problems in diverse ways.

Embodiment and feeling

They explore embodiment as foundational to cognition, tracing influential work from Antonio Damasio and Mark Soares on how bodily states influence decisions and feelings. The embodied view challenges mind in a vat scenarios and raises questions about whether embodied AI could reach a form of consciousness, or whether true consciousness requires a living, vulnerable body.

AI, brains and machines

The conversation turns to AI and the idea of computational functionalism, debating whether consciousness could emerge by running the right software on any substrate. Chalmers-like thought experiments about substituting neurons with transistors are discussed, along with Bard-like AI models and the limits of current neural architectures. The embodied cognition perspective is argued to complicate simple software-only accounts of consciousness.

Ethics and future directions

Pollan and Seth touch on the moral implications of consciousness, especially in relation to AI, psychedelics, and the ethical status of plants. The potential value and risks of pursuing conscious AI are weighed, with Pollan skeptical and Seth urging humility about what we can claim to know. They also reflect on how psychedelic experiences influence scientific understanding of consciousness and the boundaries between subjective and objective knowledge.

Live audience interaction

Questions from the theatre and livestream broaden the discussion, prompting reflections on how to measure consciousness, the role of reports from conscious beings, and the responsibilities of researchers in communicating complex ideas to the public.