Consciousness, Reasoning and the Philosophy of AI with Murray Shanahan

This episode explores the philosophical implications and evolving understanding of artificial intelligence, featuring Murray Shanahan, Professor of Cognitive Robotics, who shares his insights on AI's progress and its impact on our perception of intelligence and consciousness. Shanahan reflects on the predictive accuracy of science fiction films like "Ex Machina" and "Her," noting how the latter anticipated the current prevalence of relationships with disembodied AI systems. The conversation transitions to the historical context of AI, including Shanahan's interactions with John McCarthy, the pioneer who coined the term "artificial intelligence," further discussing the shift from symbolic AI to neural networks and the resurgence of symbolic techniques in modern language models. As the discussion pivots to reasoning in AI, Shanahan differentiates between computer science's formal logic approach and the more flexible reasoning exhibited by large language models, questioning whether AI's reasoning is genuine or merely mimicking patterns from training data. More significantly, the conversation tackles the complexities of testing AI capabilities, contrasting the traditional Turing test with the "Garland test" from "Ex Machina," which challenges individuals to recognize consciousness in a known robot. The episode concludes with reflections on anthropomorphizing AI and the need for a nuanced vocabulary to describe these "exotic mind-like entities," emphasizing the importance of considering the potential for suffering in AI development and the evolving ethical considerations.