This single author book, first published in July 2024, is intended for the general reader interested in artificial intelligence and its similarities to human intelligence. Its available to buy from many stores and also direct from Routledge (£12.99 paperback) or Amazon (£12.21 paperback, £8.96 Kindle). Please feel free to leave comments about the content and ideas in the book at the end of this blog.

Artificial intelligence is rarely out of the news and the prospect of creating entities that might match or even surpass human intelligence both intrigues and frightens us. However, expert opinion is split. Some authorities argue that superintelligent machines are just around the corner, and that we are right to be concerned, others say that current AIs only mimic intelligence rather being truly intelligent themselves.

I have been studying intelligence for all of my adult life both to understand what human intelligence is and to ask whether it can be replicated in machines. My research journey has included experimental work on the brain and the networks of nerve cells that make up our nervous systems, alongside programming computer models of the brain that look to understand how the brain works in functional and practical terms. To further explore the potential of machine intelligence I have also worked with robots since having a robotic body gives AI the capacity to sense and act in the real world just as people do.

This background provides me with a relatively rare, even-handed (I hope) and close-up view of both how intelligence has come about in the natural world, and how modern technology could create synthetic systems that could think like us. This is the reason for this book. I want to communicate this understanding to a wider audience and to show that natural and artificial intelligence are both similar and different in interesting ways.

The “Psychology of Artificial Intelligence” looks at the different kinds of human intelligence and asks if intelligence is really one thing or many. It then looks at progress in AI from its earliest beginnings through to the most recent “deep” neural networks and large language models. The book argues that AIs should be seen as genuinely intelligent but not yet capturing all aspects of human intelligence. The potential for AI to surpass human intelligence is seen as both a risk but also as an opportunity to advance human intelligence and to improve our understanding of ourselves.

Contents

Here’s a brief summary of what you’ll find in the book:

Chapter 1: Introduction

This chapter provides a definition of AI as the “the science and engineering of making intelligent machines” and discusses the multiple ways in which AI is related to psychology defined as “the science of understanding human intelligence”.  These include the shared subject matter of understanding the nature of intelligence, the use of AI methods to create and test theories of human intelligence, and the use of psychological approaches to explore how people see and behave towards AI. 

Chapter 2: What is intelligence?

This chapter explores the nature of intelligence so that we can have a clearer idea of what it might meant to create intelligent artefacts. The early history of AI is traced from its origins in cybernetics and the Turing test is discussed as a means of measuring success in creating AI. Next, the origins of psychological theories of intelligence are traced beginning with Aristotle’s distinction between perception and intellect. I then explore what we have learned about intelligence through IQ testing and introduce the question of whether humans have a single general intelligence or multiple specialised intelligences. 

Chapter 3: Brains and computers

This chapter explores similarities and differences between brains and computers.  The example of a personal digital assistant is used to examine the different technologies that must come together to make contemporary smart devices such as mobile phones as intelligent as they are (in my view, they are really quite intelligent already).  The human brain is described with particular focus on neurons (brain cells) viewed as processors, on their organisation in neural networks, and on the patterns of electrical signals that neurons use to share information. The chapter then turns to how brains give rise to minds and thence to intelligence.  Key principles of brain architecture are introduced as well as the different types of learning that are observed in brains. 

Chapter 4: The building blocks of intelligence

This chapter explores the building blocks of intelligent systems, both natural and artificial, from different kinds of reasoning and control, through to processes such as search and optimisation. It also explores some of the reasons why creating AI is so challenging, discussing the “combinatorial explosion” that makes search costly and time-consuming and the “frame problem” (knowing what is relevant) in reasoning. The chapter also asks whether humans think differently from machines, exploring several unsolved challenges, including analogical thinking and abductive reasoning, and looking at the role of emotions in human decision-making.

Chapter 5: Learning in neural networks

This chapter explores brain-style computation in AI in the form of artificial neural networks (AIs composed of very large numbers of highly interconnected but relatively simple processors) and asks whether today’s “deep” networks have important similarities to the human brain (I argue that they do).  The chapter also examines the “transformer” architecture of recent large language models (LLMs) such as ChatGPT and their capacity to generate fluent and original linguistic output. I point out that although LLMs are very large (in terms of the number of processors and connections) they are still small compared to human brains (<1% of the size of the human connectome), although they are approaching the size of some of the language areas of the brain.

Chapter 6: Towards artificial general intelligence

In this chapter I discuss AI’s capacity to understand the world in relation to a theory of meaning as comprising sense—how words relate to other words—and reference—how words relate to the external world.  I argue that LLMs already encode meaning in relation to sense but not in relation to reference.  To obtain a better grasp of meaning, AIs will require a greater capacity for unmediated two-way interaction with the world through robotic bodies. I also discuss the broader class of generative AIs models (AIs that learn to generate new data that could belong to the data-set on which they were trained) in relation to theories of predictive processing in the human brain. The chapter concludes by arguing for a nuanced view of multiple intelligences, including that humans combine fast, pattern-recognition intelligence with a slower sequential reasoning capacity, and by arguing that humans can be considered to belong to the class of “cyber-physical systems” that includes complex, embodied AIs such as robots.

Chapter 7: Living with artificial intelligence

The final chapter considers how people can live alongside AI, including our attitudes towards AI and our possible relationships with it. The chapter also looks at some of the ethical and societal risks of AI, including error and bias in AI algorithms, the risk of “artificial stupidity” (putting too much trust in AIs that have limited understanding of the world they are in), and the consequences for our long-term future of an AI “singularity”—the point in time at which AIs become superintelligent.  I argue that, if developed with care, AIs could help to address some of the most pressing concerns in the world today.  The book concludes that the partnership between psychology and AI has the potential to unlock further insights into both natural and artificial intelligence and could help to answer some central questions about what it means to be human.

Reviews

Review from “The Psychologist”: “The book manages to maintain a balanced perspective, highlighting both the transformative potentials of AI and the risks it poses. I enjoyed it not just for its technical discourse but also the philosophical reflection on how AI reshapes our perceptions of identity, intelligence, and the future of human evolution. Reading this book led me to reevaluate my own perspectives on technology and humanity; I feel both enlightened and slightly alarmed thinking about the real-world implications of AI that Prescott lays out so vividly.” Review by Zilong Zhong, 12th November 2024.

Press Coverage

The Guardian. Could AI cure the downward spiral of human loneliness?

Sify.com. AI and loneliness: Culprit or cure?