Richard Harper, Emeritus Professor in the School of Computing and Communications, explores how our understanding of intelligence – and a broad inclusive view of what intelligent might be – can shape the way we use AI technologies for the better in his critically-acclaimed new book, “The Shape of Thought: Reasoning in the Age of AI”.

The book – Professor Harper’s final before his retirement – was a labour of love that took two years to complete and was published in October by McGill-Queen’s University Press. “The Shape of Thought” grapples with the notion that the recent resurgence of AI technologies has witnessed a “narrowing” of our definitions of what constitutes intelligence, reducing it to something mathematically calculable with tangible and observable outputs (and therefore computationally reproducible), as opposed to its true multifaceted and inherently “human” nature, shaped by morals and moods, and with the capacity for innovation.

Professor Harper begins his endeavour with a brief chronology of the history of chatbots and AI, suggesting that many of these systems (both historical and contemporary), are simply advanced “pattern recognition” engines, capable of successfully identifying recurring themes, but limited hugely by the domains in which they were designed for use and the data they have been trained on. While incredibly useful for many tasks, are not indicative of all the forms of intelligence that can be found, he says.

He argues that chatbots such as ChatGPT are better off viewed as “word geometry” engines rather than something approaching “artificial intelligence”, tools that can be used to aid our understanding and communication but not something that can produce meaning on our behalf. The algorithms on social media that use “intelligence” to suggest videos, posts, and friends might better be termed “proximity engines”, utilising commonalities between established connections and communities to suggest similar content, but not the intelligent assistant many of these sites would lead us to believe.

Throughout “The Shape of Thought”, Professor Harper emphasises that intelligence and reasoning should not be considered without the context of their usage, and that systems that may seem impressive in one situation falter in others where humans – with our ability to adapt, capacity for critical thinking, and wealth of actual lived experiences to draw upon – out perform machines time and time again.

Critical to Professor Harper’s thesis is his emphasis that we should reconfigure our approaches to so-called “AI”, and approach it as an aid to making us, as users, more intelligent or thoughtful rather than being in possession of intelligence itself. Rather than trying to emulate human cognition, Harper suggests AI engineers should focus instead on ways in which their tools can augment current and future human experience, considering the ways in which we work and live and embedding into those existing infrastructures in order to enhance our practices rather than looking to replace human input.

“The Shape of Thought” has been well-received by critics from across both academia and industry. Professor Alan Blackwell from the University of Cambridge states that ““The Shape of Though” cements [Harper’s] reputation as one of the leading philosophers helping us to understand everyday technologies, now and for the future”, whilst Antoine Tardif, founder of Unite.AI describes it as “not just a critique of the present”, but as “a roadmap for how the next generation of AI systems might be conceived—grounded, contextual, and aligned with the realities of human thought rather than abstract fantasies of machine intelligence.”

On the reception of his book, Professor Harper commented: “I was delighted to receive such wonderful comments about “The Shape of Thought”. You work hard on a book but don't really know how a book will be received - you only learn once the reviews arrive. Until then, it’s a prayer and hope!”

“The Shape of Thought” is available for purchase from Amazon, TG Jones, and other online retailers.