Book review: “Machines behave badly”

Academic Toby Walsh dissects the idiosyncrasies of the community developing AI and reshaping our world, with a self-reflective insider perspective.

Books on the societal impacts of artificial intelligence and the ethical questions it raises are proliferating, so it’s always worth asking: is this book doing anything different? In the case of “Machines misbehave: the morality of AI” (The History Press, £20, ISBN 9780750999366), the answer is yes, thanks to the author’s decades of grounding in the community where AI originated.

The first third or so of “Machines Behaving Badly” looks at people and companies developing AI technologies. This is packed with little-known nuggets of information. It’s unclear, for example, how small the group of people building AI is (“There may never have been a planetary revolution before that was led by such a small group of people”).

Walsh portrays this community with more than general traits and statistics. He speaks in revealing detail about uncomfortable sexism and community-based philosophies: objectivism, techno-libertarianism, and transhumanism. In the most thought-provoking part of “Machines Behaving Badly”, he discusses the influence of Ayn Rand’s philosophy on Silicon Valley and how it influences the construction of AI: “A robot has a contact direct with the reality of the world via his perception of this world, his only goal is to maximize his reward function, and he does this by reasoning rationally on these precepts.

He then vivisects AI companies: their cheap money, their intent to disrupt, how the voting rights attached to Big Tech stocks tend to liberate executives from accountability. Importantly, Walsh offers alternatives to this dominant model.

The rest of the book covers more familiar ground. We find the same case studies (micro-targeting of social media for political purposes, racial bias in facial recognition software) and debates that we find in dozens of other books. However, Walsh uses accessible language and explanations suitable for readers unfamiliar with AI and its unfortunate effects. He does a particularly thorough job of dissecting the issue of lethal autonomous weapons, having been a key organizer in the campaign to halt the development of these technologies.

In a distinctive twist, an entire chapter is dedicated to AI and climate; it’s encouraging to see this topic getting the same prominence as better-trodden AI issues.

Walsh leaves the reader with some important lessons, particularly about the limits of AI. If only more people recognized how distant AI is from human intelligence, he writes, we could see fewer road deaths caused by drivers who place too much faith in the algorithms of self-driving cars.

The purpose of the book is, according to Walsh, “to open your eyes to this strange intruder, to make you think about the unintended consequences of AI.” Machines Behaving Badly is a good introduction to the unintended consequences of AI and contains some genuinely thought-provoking insight into the quirks of the people behind them.

Sign up for the E&T News email to get great stories like this delivered to your inbox every day.

James G. Williams