Interviews with Minds


Who are all these people writing chapters in this book?

In spring 2000, the major UK society for Artificial Intelligence (SSAISB) held its annual spring conference at the University of Birmingham. One of the symposia (led by Professor Aaron Sloman of the University of Birmingham) had as its central theme "How to Design a Functioning Mind". Aaron could not find the time to collate an updated and extended version of the presentations as an edited text, so I took the task on. The authors present at the symposium, from across the world (well at least from Jerusalem to the west coast of the USA), plus others working in the field have updated their perspectives on this and related questions in the chapters they contribute to this text.

What is important about this book?

This book brings together a number of disparate perspectives on questions such as "What is Mind?","What kind of things can have minds?", "What are the qualities associated with mind?" and "Can we build synthetic or artificial minds?". Rather than present a single argument for any one of many alternative and sometimes intersecting perspectives, the text presents a collection of coherent arguments that aim to foster unblinkered thought on these and related questions.

Surely only humans (and perhaps some other higher order animals) have minds?

The question presumes an answer to a more fundamental question such as "What is mind?" There are many qualities that can be associated with mind. Some such as reasoning and the expression of that process via natural language preclude all but a very small set of living creatures. One definition used within the text describes mind as that which enables an agent (whether living, biological or completely synthetic) to decide what to do next. Other definitions of mind dwell on qualities such as free will, consciousness and that mind is used for fruitful social and environmental interaction. Such a definition of mind that focuses on the last quality does not exclude ants or machines.

Machines cannot have minds, they are just a collection of metal, plastic and semi-conductors.

The question again presumes an answer to a more fundamental question such as "What is mind?" There are qualities associated with mind that are today implemented or theoretically implementable on current computers. Over the history of computation and artificial intelligence, it has seemed that the qualities required to demonstrate the type of intelligence associated with mind shift as they become implementable on computers. One example used to be, for example, Chess. Chess playing computers can beat most players, even a world grand master can be beaten. Computers fly civil aircraft more effectively than many trained pilots. Computers can sort addressed mail faster and more effectively than trained postman. On the other hand a child is better at interacting with its world than any current computer. An ant colony is better organized than any current computer network. From a different perspective the question presumes that mind cannot exist without a living body associated with it. Chapters within the text address this. A final thought on this question is that while the vast majority of current computers are made from metal, plastic and silicon, not all current computers are and certainly in the future we can expect many different types of materials to be used to support many different types of computation.

Why do game playing machines need minds?

Why do children, animals (and even grown-up people) play? For many different reasons including social interaction, learning, skill acquisition and development and sometimes just for fun! Current computer based entertainment (whether game playing or otherwise) relies heavily on the human user(s) to inject the fun and keep the interactions interesting. A game playing machine should want to play for the same reasons (or at least some of the same reasons) as a child. And you need a mind to have fun, even if it is mindless-fun!

What has emotion to do with an artificial mind?

There is a growing realisation that designing an artificial mind without emotion, or affect if you want to avoid the emotive term emotion, is impossible. Affect is highly correlated to motivation and social interaction. Affect (and emotion) are not just associated with "being emotional" but are fuindamental to decision making and actions in an environment. Neuroscientists, psychologists, cognitive scientists and practitioners of artificial intelligence are interested in the role of affect and how it relates to decision-making and behaviour. Many of the chapters in this text have something to say on why or how emotion should be incorporated into the theory, design and engineering of artificial minds.

What has social science to do with an artificial mind?

There is a growing realization within certain sections of the artificial intelligence and cognitive science community that social psychology and social studies in general have important things to say about the processes and qualities that are associated with the study of mind. If situated cognition is correct (to any degree at all) then much of most animals environmental actions are about interaction with species similar or species dissimilar agents. Language is about social interaction. Language involves the manipulation of multi-media and symbols. Computers can do the latter two effectively in certain domains; less so than in other domains. If minds are about communication and interaction in social environments, then social studies have something to say about the nature of artificial minds.

What has chemistry to do with Artificial Intelligence?

Traditional models of computation (whether arithmetic, algorithms, symbol processing or most neural networks) work, at some level of abstraction, in terms of a Turing Machine. Biological brains are not. They are asynchronous evolved structures that do not fit within the confines a Turing Machine model. Biological brains also live in soup formed, in part, of control chemicals, energy and oxygen Biological brains can globally change state when the chemical soup changes in chemical or physical nomenclature. Computational chemistry attempts to model the processing such chemical systems. Computational chemistry can be used in investigating computational models of the mind too. Perhaps future computer hardware will include a computational equivalent to the chemistry underlying the functioning of the mind. There are chapters in the text that discuss this. Quantum computation is also included.

Do you mean you are building the machines that will control the future?

This is a strange question. At the start of the twenty-first century so much of western (and far eastern) civilization relies on the machine whether digital or analogue that the question is pointless. Even a pocket-calculator can perform mathematics faster (and more accurately) than even the most gifted of mathematicians. However, there is no reason to suggest that pocket calculators are going to control the future. They are simply part of it. Some of the chapters allude to sociological norms and comprehensions of what is involved in machine intelligence, and relate their arguments to instances pulled from popular culture. However, the ethics of creating intelligent agents is a bonafide research topic. To what extent scientists take responsibility for the uses of their pure or even applied research when it is put to unexpected uses is a laid query. The ancient Chinese genius who discovered how to make gunpowder enabled brilliant firework displays and a more effective quarrying of rock. He also enabled the musket and the gun. The user of technology is the one responsible for the uses that technology is put.


Copyright: D.N.Davis Last updated: 25 November 2003