Murray Shanahan: AI Consciousness, Cognition, and Human Interaction with LLMs

Clip title: Consciousness, reasoning and the philosophy of AI | Murray Shanahan Author / channel: Google DeepMind URL: https://www.youtube.com/watch?v=v1Py_hWcmkU

Summary

This podcast episode features Professor Murray Shanahan, a Professor of Cognitive Robotics at Imperial College London and a Principal Research Scientist at Google DeepMind. Shanahan, known for his work in artificial intelligence and as a consultant for the film Ex Machina, delves into the profound philosophical questions raised by the rapid advancements in AI, particularly Large Language Models (LLMs). The discussion explores the nature of mind, consciousness, reasoning, and the ethical implications of how humans interact with and perceive increasingly sophisticated AI systems.

Shanahan highlights the predictive accuracy of the film Her (2013) over Ex Machina (2014) regarding the human relationship with AI. While Ex Machina provocatively questioned AI consciousness, Her foresaw the development of powerful, disembodied language models with whom humans could form deep, albeit artificial, emotional connections – a reality mirrored by current LLMs. He traces the history of AI, mentioning John McCarthy, who coined “Artificial Intelligence” in 1955. Shanahan prefers the term “Artificial Cognition” for its precision in describing information processing, though he acknowledges “AI” has become ingrained. He explains that early “Symbolic AI” focused on explicit rule-based reasoning, which proved brittle and difficult to scale due to the challenge of codifying vast amounts of common-sense knowledge. This paradigm, which he felt was “doomed” by the early 2000s, has now seen a resurgence in new forms with LLMs, particularly in “chain-of-thought” reasoning, where models generate logical steps to arrive at an answer.

The conversation then tackles definitions of intelligence and consciousness, and how we test for them. Shanahan dismisses the Turing Test as a measure of true intelligence, arguing it’s too narrow and easily passed by today’s LLMs which excel at linguistic mimicry. He proposes what he calls the “Garland Test” (inspired by Ex Machina), which asks if one still attributes consciousness to an entity knowing it is a robot. For Shanahan, consciousness is a multifaceted concept, encompassing world awareness (which LLMs lack due to their disembodied nature), self-awareness (where LLMs show nascent reflective capabilities), and sentience or the capacity to feel/suffer. He emphasizes that the question “Can AI be conscious?” is flawed, as humans tend to bundle these aspects of consciousness, while non-human animals and potentially AI might exhibit them separately. He stresses the ethical implications, particularly around suffering, which he believes current AI, lacking a physical body, cannot experience.

Ultimately, Shanahan calls for a new conceptual framework and vocabulary to describe AI. He refers to LLMs as “exotic mind-like entities” – “mind-like” with a hyphen to denote their similarity without fully equating them to human minds, and “exotic” to highlight their fundamental difference, such as being disembodied and potentially having unique forms of selfhood. He suggests that while we must be cautious about “bad anthropomorphism” (misleading emotional attachments), “good anthropomorphism” (treating AI politely or attributing intentions) can be useful for interaction, as it mimics natural human tendencies. The increasing sophistication of AI will inevitably lead to new ways of understanding and interacting with these entities, pushing the boundaries of our existing linguistic and philosophical concepts in ways we cannot yet fully predict.

Video Description & Links

Description

In this episode, Hannah is once again joined by Murray Shanahan, Professor of Cognitive Robotics at Imperial College London and Principal Scientist at Google DeepMind, for a philosophical deep dive on AI. They explore everything from consciousness and metacognition in animals, to symbolic AI and neural networks. Murray also shares insights into his involvement with the film ‘Ex Machina’ and discusses the idea of reasoning, anthropomorphism, and the future of AI.

Further reading / listening: Toward the future, S1 Ep 7: https://www.youtube.com/watch?v=yf31XT1G1RQ Papers by Murray: https://scholar.google.co.uk/citations?user=00bnGpAAAAAJ&hl=en

Timecodes 00:00 Introduction 02:00 Ex Machina 03:28 John McCarthy & Coining “AI” 05:45 Symbolic AI 09:43 AI Reasoning 14:42 Turing Test 18:10 Alternative Tests 21:58 Anthropomorphism 26:42 Embodied AI 28:17 AI Consciousness 34:00 Octopus Analogy 37:48 Prompt Engineering 39:40 Exotic Mind-Like Entities 40:44 Conclusion

Presenter: Professor Hannah Fry Series Producer: Dan Hardoon Editor: Rami Tzabar Commissioner & Producer: Emma Yousif Music composition: Eleni Shaw Audio engineer: Richard Courtice

Production Manager: Dan Lazard Video Studio Production: Nicholas Duke Video Director: Bernardo Resende Video Editor: Alex Baro Cayetano, Bilal Merhi Audio Engineer: Perry Rogantin Camera and Lighting Operator: Robert Messere Production Coordination: Zoey Roberts, Sarah Ellen Morton Visual Identity and Design: Rob Ashley Commissioned by Google DeepMind

NemoClaw Knowledge Wiki

Explorer

Murray Shanahan: AI Consciousness, Cognition, and Human Interaction with LLMs