Margaret Boden – Creativity and Artificial Intelligence – Interview
In this interview, Margaret Boden, Research Professor of Cognitive Science at the University of Sussex, explores the intersection of human psychology, creativity, artificial intelligence (AI), and neuroscience. She emphasises the importance of viewing the mind’s functions in computational terms to demystify complex cognitive abilities like creativity, arguing that creativity is not magic, but a process that computational concepts can help us understand.
Boden points out that current neuroscientific methods – like brain imaging – are limited because they largely show biological correlations (like associative memory or blood flow) without explaining the deeper “why” or “how” of complex thought processes, such as evaluating the relevance of creative insights or engaging in hierarchical thinking. She stresses that neuroscience needs to be guided by structural theories from psychology and AI to ask the right questions. She highlights the work of Uta and Chris Frith on autism and “theory of mind” as a prime example of psychological theory successfully guiding neuroscience.
Finally, Boden shares her skepticism regarding AI’s current trajectory toward general, human-level intelligence. She critiques systems like IBM’s Watson, noting that their achievements stem primarily from raw processing speed and statistical data associations rather than genuine comprehension. She argues that to truly reach human-level intelligence, an AI must be capable of deeply understanding the nuance, subtlety, and metaphor of human language – such as grasping complex poetic imagery like Shakespeare – something current systems are far from achieving.
Chapters
0:00 Introduction to Cognitive Science & the Computational Mind
1:00 Demystifying Creativity
2:04 The Current Limitations of Neuroscience in AI
3:48 Connectionism vs. Symbolic AI
5:42 Hierarchical Thinking and Brain Functions
8:30 The Need for Synergy Between Psychology, AI, and Neuroscience
11:29 Case Study: Autism, Theory of Mind, and Brain Scanning
18:25 Is General Human-Level AI Actually Possible?
19:42 Critiquing IBM’s Watson and Associative Thinking
23:28 The Ultimate Test: Language, Poetry, and True Comprehension
Also I video recorded Maggie’s talk at AGI12 / Winter Intelligence: https://www.scifuture.org/why-computers-arent-writing-shakespeare-yet-reflections-on-margaret-bodens-agi12-talk/
