The machines may not win, after all
IT IS POSSIBLE TO argue that people involved in the minutiae of a particular field are among the least likely to provide a cogent analysis of its broader implications.
A mechanical engineer whose day job involves seeking incremental improvements in high-performance motorcars, for instance, might not be a credible source regarding challenges to transport infrastructure.
This potential for macrocosmic opacity to arise from microcosmic clarity is nowhere more present than within the area of artificial intelligence. Thus, when a book written by an eminent professor in the field cites inaccurate information about AI relayed by “people, especially those outside the field”, and informs the reader, via footnote, that the book “is a moment where I look up from my work”, fears of a text too narrowly focused seem reasonable.
Thankfully, however, in the case of It’s Alive by Toby Walsh, professor of AI at the University of New South Wales, such concerns are quickly dispelled.
Walsh delivers an accessible and concise history of artificial intelligence, with Alan Turing as its fulcrum but stretching back to Gottfried Leibniz. He follows with a longer exploration of the current state of play within the discipline, and then ends with a series of cautious predictions about the near and more distant future.
Three things need to be said by way of global observations here. First, Walsh delves into the technicalities of his world only sufficiently to provide the detail his readers require to comprehend his points.
Second, he is adept at carefully treading a middle path, steering clear of the technoutopianism that glosses over the manifest challenges presented by AI, on one hand, and the doom-laden keening of pessimistic futurists on the other. Third, Walsh’s ability to achieve what he sets out to do with the first two points means that this book, with text covering just 294 pages, could easily have been much longer without either baffling or boring the reader.
In some ways Walsh is at his most interesting when his focus is at its widest. In this mode, he provides informative insight into some AI developments rolling out now, and others that remain merely promises – or nightmares, depending on your perspective.
When addressing the advent of autonomous vehicles, he is refreshingly free of the gung-ho free-market determinism that runs through much of the discourse surrounding the brave new world of artificial intelligence and robotics.
Rather than seeing the development of driverless taxis and robo-trucks as an exciting and inevitable consequence of Silicon Valley laissez faire, he calls for a strong regulatory role for government.
“We do not let drug companies test their products freely on the general public,” he states. “Similarly, we should not let technology companies test autonomous cars freely on public roads without strong oversight.”
(He is even sterner when it comes to autonomous weapons, and in 2015 helped draft an open letter calling for a permanent ban on their development. Today that letter contains 20,000 signatures, including those of Stephen Hawking, Elon Musk and Noam Chomsky.)
One of the companies lunging towards an autonomous future is the crowd-sharing ride service Uber. Being an Uber driver, comments Walsh dryly, is one of the world’s newest jobs, and is also very likely to be one of the shortest-lived.
The widespread loss of employment – and whole employment categories – is a subject frequently raised by technopessimists, such as the futurist Noah Yuval Harari. There are good reasons for this, given pretty much everyone in the field agrees job losses will be massive. Walsh, however, attempts to bring to calm to the subject by reviewing the findings of a key 2013 Oxford University report that broke down likely redundancies sector by sector.
He has good and bad news. Creatives, such as novelists, will be largely safe, he says. The prediction that bicycle repair will be almost completely automated is “rubbish”, because bicycles are fiddly, bike