Book Review: Artificial Intelligence: A Guide for Thinking Humans

Mitchell, Melanie. Artificial Intelligence: A Guide for Thinking Humans. New York, NY: Picador. 2019.

Though well into its sixth year of publication, Melanie Mitchell’s 2019 work on artificial intelligence remains relevant and instructive.

Her approach appeals to a broad audience, speaking to the noob and to the technically savvy. (The author of this review plants herself firmly with the former.) In fact, Michell’s book comes highly recommended by AI itself, being listed as a quintessential work in Chat GPT’s list of recommended reading on AI for beginners.

With an impressive CV in computer science and AI, the author takes the reader from what seems like ancient history (the mid 20^th century) into the unknown future, working her way from the earliest iterations of symbol manipulators to the complexities of play and language and the frustration of seemingly simple barriers.

The reader lacking technical expertise will find many passages challenging, but the book is accessible in its presentation of the dream vs. reality of AI and the work needed to perfect it and create guardrails for its effective use.

The AI optimist will be delighted in the infinite applications of AI, from medical diagnostics to gaming, but most readers likely identify as skeptics. Mitchell reveals the reasons why so many inherently distrust this technology: even those who created it do not fully understand how it works. And those who understand enough can manipulate it to myriad nefarious ends.

Even so, there are moments of levity when sophisticated AI makes mistakes and it becomes clear, as Mitchell notes repeatedly, that simple human things are hard for machines. Creating machines with common sense and therefore real human intelligence appears to remain a distant prospect.

In the meantime, there are major questions to answer about the ethics and regulations we need to keep AI from misdiagnosing diseases because it did not learn enough varieties of skin pigmentation; from falling prey to adversarial and criminal attacks; and from omitting just one possible scenario that would hurt someone in an AI driven car.

This book provides an excellent overview of AI but reaches only one conclusion: the age of AI is a remarkable opportunity, and every thinking human must be engaged.