(2016) Jerry Kaplan, Oxford University Press, £10.99 / US$16.95, pbk, xvi + 164pp, ISBN 978-0-190-60239-0
Dr. Jerry Kaplan teaches at Stanford University, is a Fellow at the Stanford Center for Legal Informatics, and has co-founded four silicon valley start-ups (two of which became publicly traded companies). His specialist subject is Artificial Intelligence (AI). This all means that he should know a lot about the subject at hand and, indeed, this book has a lot of interesting things to say and offers a lot to think about.
The first chapter, ‘Defining Artificial Intelligence’, asks the obvious question - what exactly is Artificial Intelligence? His answer is that there is no exact definition and then goes on to look at the differences between human intelligence and a well-programmed machine. He concludes that sometimes, in some circumstances or for certain jobs, there may be little difference or even that the right machine might do a better job.
Moving on to ‘The Intellectual History of Artificial Intelligence’ he describes the history of AI, starting with the term’s first use in 1956 by John McCarthy (an assistant professor of mathematics at Dartmouth College in Hanover, New Hampshire). This includes the rise (and fall) of expert systems, machine learning, and artificial neural networks. Next, in ‘Frontiers of Artificial Intelligence’, he discusses robotics, computer vision, speech recognition, and the processing of natural language. He then considers the ‘Philosophy of Artificial Intelligence’ and asks whether a computer can think, have free will, be conscious, or have feelings, as we would define such terms.
In the 5th chapter, ‘Artificial Intelligence and the Law’, Kaplan explains how AI is already helping lawyers, how it might affect the practice of law, and changes that may be required to the law. He also discusses the legal position when AIs or automated systems create or agree contracts and asks if they can own property, commit crimes, or be held legally accountable, as per corporations. Moving on to ‘The Impact of Artificial Intelligence on Labour’ he looks at automation and the impact that it has on jobs, concluding that advances in technology have always caused jobs to disappear but that others come along and, as the effect is usually spread over a long period, the long-term effects are naturally ameliorated. For example in the United States: in 1870 some 70-80% of the labour force was engaged in agriculture but by 2008 this had dropped to less than 2%; this was hardly cataclysmic as the labour market adapted over the 140 years. However, the advances in AI and automation are likely to be at a pace hitherto unseen and the impact on jobs (and thus society) could be very serious unless steps are taken as part of the process. At present, it looks like 47% of jobs are at high risk of automation and a further 19% at medium risk over the next couple of decades or so. He provides interesting tables on the jobs currently at greatest and at least risk; for example, machine operators and clerical workers are at high risk whereas medical practitioners, systems analysts, and programmers are hard to automate.
In ‘The Impact of Artificial Intelligence on Social Equality’ he discusses who will benefit from these technological advances and how social disruption can be minimised. He looks at supply and demand and asks if we need a prosperous middle class or could we manage with a rich class who spend their copious money having everybody else work for them and cater to their whims. He wonders if we could have an economy based on sizable government handouts in lieu of wages and simply accept the idea that few will even be able to find work. The final chapter, ‘Possible Future Impacts of Artificial Intelligence’ covers the rate at which various aspects of AI and associated systems are changing and examines the ‘singularity’, the point at which AIs can improve themselves. He asks if runaway super-intelligence is a real concern and concludes that it is not. On the other hand, he warns that there is a real danger of intelligent systems getting loose, citing viruses and the like that have already ‘escaped’ their creators. He finishes by looking at the possibility of uploading oneself, or one’s memories and attitudes, into a machine and asks whether that would be really be you.
Throughout this has been both light but thoughtful; the reader is not overwhelmed by long technical arguments yet is given plenty to think about. I found some of his points in the chapter on the Philosophy of AI a little difficult to follow and he certainly did not carry me with him, but maybe that is just my lack of experience in this aspect of the field - or maybe it is just philosophy in general! All told, a light humour was apparent in some of the examples he cited during his discussions yet his points were very serious. I found this to be well worth reading; it covers serious matters from an intelligent point of view and leaves one with much food for thought.
Peter Tyers
[Up: Non-Fiction Index | Top: Concatenation]