In the history of mankind’s scientific endeavours, perhaps no other advancement has been talked about as much before it actually came to pass as artificial intelligence (AI). The subject has provided rich fodder for writers of science fiction books and filmmakers, who have been speculating about the nature of the technology for well over half a century.

And now, with the culmination of decades of progress in building advanced computing devices and machine learning putting the world on the cusp of true AI, the ground is finally fertile for philosophers, scientists and all manner of other experts and thinkers to jump into the discussion with their own vision of the future of the thinking machine.

What to Think About Machines That Think is a compilation of incredibly short essays on AI, edited by John Brockman, the not-so-well-known literary agent to some of the finest minds of our times.

Brockman has his roots in the same avant-garde Greenwich Village scene in New York of the 1970s that nurtured the genius of creative thinkers such as Bob Dylan and Andy Warhol.

In the mid-1990s, Brockman famously proposed the concept of the ‘Third Culture’, referring to the emergence of a new breed of intellectuals who were communicating the scientific message directly to the masses, cutting out the traditional men of letters, as science increasingly became a part of popular culture.

Brockman currently runs the ‘online science salon’ edge.org (which The Guardian once called “the world’s smartest website”). The Edge Question, which he poses every year to his extended network of exceedingly intelligent friends and clients, is a simple, direct question that seeks to push the boundaries of understanding on a burning scientific issue.

Year 2015’s question prompts respondents to ruminate on the potential technological, ethical and even emotional issues that will arise when the first machines start to think independently.

Future glancing The book is ostensibly an attempt at peering into the future by standing on the shoulders of giants in the fields of cognitive science, psychology, machine learning, physics, robotics, cybernetics, philosophy, economic theory, anthropology, neuroscience and several other sciences, all of which have a bearing on the development of AI.

The essays have no particular format, but are — for the most part — concise and highly accessible. The style of writing changes constantly — a given when a book features nearly 200 different authors — and the end result is only engaging some of the time.

Some essays, written by experts in fields that are at best tangentially related to AI such as the musician Brian Eno or the ad-man Rory Sutherland, feel overly speculative and superfluous.

Meanwhile, those by authority figures such as philosopher Nick Bostrom, whose Superintelligence is one of the most influential books on AI in recent times and Harvard psychologist and cognitive scientist, Steven Pinker seem to have lacked depth and could have been a lot longer than they were.

The compilation does not make for particularly easy reading. Despite the plethora of PhDs, published literature and sundry qualifications that back them up, it is easy for a casual reader to wonder whether any of the contributors featured in the book actually know what they’re talking about.

Biases and intelligence AI is a field that is still in a nascent stage of development and its importance to human civilisation means that conversations regarding its implications have begun very early.

This also means that talking about artificial general intelligence — a machine that can really think for itself, at this stage almost certainly involves a certain degree of gazing into a crystal ball that is clouded by personal biases.

Within the first 50 essays, readers will find out how human level AI is within touching distance; how it is still just a pipedream; how the machines are going to murder us all; how they’re going to benevolently rule over us; how they will leave us to rot on Earth while colonising galaxies; and how the future will almost certainly see human and machine intelligences merging.

For a book which — by decree of Brockman — religiously avoids references to sci-fi cannon, it features rather a lot of outlandish plots that a struggling science fiction author could very easily draw inspiration from.

The most important purpose of this book will be its ability to serve as historical record of the confusing maelstrom of conflicting theories that represents the best set of opinions on artificial intelligence available at this time.

For the pros For the layman, who has no pre-existing frame of reference regarding artificial intelligence, this book could be a thoroughly disorienting experience that could leave them questioning their own intelligence.

For the AI enthusiast who has already chosen a side, it will provide additional armour with which to fortify existing beliefs and plenty to disagree with.

However, unless you are well and truly wedded to your beliefs, it is hard to hold on to them through a thorough perusal of this book — the constant back and forth serving to highlight the speculative nature of everybody’s assertions.

“I do not believe that machines that think exist, or that they are likely to exist in the foreseeable future. If I am wrong, as I often am, any thoughts I might have about the question are irrelevant. If I am right, then the whole question is irrelevant,” says the renowned physicist Freeman Dyson in an incredibly succinct two-line contribution that cuts through the clutter like no other.

MEET THE AUTHOR

Founder and publisher of the influential science portal Edge.org, John Brockman has edited volumes such as The Universe, Thinking, This Explains Everything, This Will Make You Smarter, Culture, What Is Your Dangerous Idea?, and What We Believe but Cannot Prove. He is the founder of the literary agency Brockman Inc.

Related Topics
comment COMMENT NOW