The Terminal Experiment: A substandard Crichton-style thriller

The Terminal Experiment by Robert J. Sawyer science fiction book reviewsThe Terminal Experiment by Robert J. Sawyer

Robert J. Sawyer is a very popular Canadian science-fiction author, with many novels under his belt and several major awards, including the 1995 Nebula Award for The Terminal Experiment, 2003 Hugo Award for Hominids, and 2006 John W. Campbell Award for Mindscan. I hadn’t read anything of his so I decided to give The Terminal Experiment a try. It’s about an engineer who creates three artificial copies of his consciousness, and one of them becomes a killer. The audiobook, by Recorded Books, is narrated by the very competent Paul Hecht, and is an easy listen. But how well does it hold up as an award winner?

I’ll freely admit I am not a big fan of “techno-thrillers” in science fiction. Generally I find this a flimsy plot device to move a mediocre story forward quickly. Usually a shadowy and sinister organization or super-villain is up to no good, and the intrepid hero and his clever sidekick and/or love interest race against time to defeat the bad guy(s) and prevent a terrible calamity. The heroes are usually are scientists, engineers, detectives, private investigators, or scholars. Sometimes they produce massive runaway best-sellers that explore the secret history of Christianity and became Hollywood blockbusters starring Tom Hanks and Audrey Tatou, flitting from one gorgeous European location to the next, usually with a trail of dead bodies left behind.

Well, The Terminal Experiment isn’t quite that bad, but it hardly breaks any new ground or provide insight into the nature of AIs and human consciousness. I recently read Greg Egan’s Permutation City, and that book dives into those ideas in such depth and complexity it was almost overwhelming. In contrast, The Terminal Experiment goes down way too easily, following a by-the-numbers thriller plot. Dr. Peter Hobson is the biomedical engineer who invents a machine that can detect brain patterns as they leave the body after death, which many interpret as proof of a human soul. After creating much hoopla in the media and religious circles, he decides with the help of his friend Sardar Muhammed, an AI programming expert (very convenient, don’t you think?) to create three AI simulations of Peter’s consciousness, in order to test some theories about the afterlife and soul.

So three simulations are created: 1) Spirit, a version of Peter in which all physical desires and urges are removed, allowing for a pure intellect unburdened by worldly concerns; 2) Ambrotos, who has all fears of aging and death removed to simulate the conditions of an immortal being; and 3) a control version of Peter, with no special modifications.

Initially all three simulations take to their existence positively, exploring the Internet of 1995 with enthusiasm and curiosity. The book really betrays its age with some very dated descriptions of “cutting-edge” technology of the nascent web, and there are numerous laughable details about information technology, etc. Over time the simulations get frustrated with their limited virtual environs and break out into the larger global IT network. They also start to develop some aggressive behavior, seemingly triggered by Peter’s subconscious feelings.

What ensures is a thoroughly unexciting thriller as they try to outsmart the simulations and prevent them from getting out of control. This idea has been done to death many times before. I found it hard to care about the characters, plot, or even the philosophical questions the book raised, not because they questions themselves are not important (they are), but due to the amateur way in which they are presented to the reader. The writing is pedestrian but unthreatening — exactly what you would expect from a “mainstream thriller.“

What The Terminal Experiment illustrates is the problem with near future techno-thrillers winning major awards like the Nebula or Hugo. While they may seem fairly innovative or cutting edge at the time, it only takes 5-10 years to make them hopelessly outdated or wrong in their predictions. Books about the far future, alternate histories, or fantasies are less likely to age badly. In 1995 it did beat John Barnes‘ Mother of Storms, Nancy KressBeggars and Choosers, Paul Park’s Celestis, Walter Jon WilliamsMetropolitan, and Gene Wolfe’s Calde of the Long Sun, but I find it hard to believe this was the best science-fiction book of that year. In the end you can never please everyone when choosing the “best” science-fiction or fantasy novel, since taste plays such a major role, but voters should consider how well a given book is likely to stand the test of time, so when someone picks up an award winner from a previous decade they can be confident it’s at least well-written and thought-provoking.

Published in 1995. An experiment has gone terribly wrong. Dr. Peter Hobson has created three electronic simulations of his own personality. One will test life after death; another, immortality. The third one is the control unit. But now all three have escaped from Hobson’s computer into the worldwide electronic matrix. And one of them is a killer. Robert Sawyer has won many awards for his science fiction, which is praised for its blend of high-tech mystery and suspenseful pacing.

FOLLOW:  Facebooktwitterrssmail  SHARE:  Facebooktwitterredditpinteresttumblrmail
If you plan to buy this book, you can support FanLit by clicking on the book cover above and buying it (and anything else) at Amazon. It costs you nothing extra, but Amazon pays us a small referral fee. Click any book cover or this link. We use this income to keep the site running. It pays for website hosting, postage for giveaways, and bookmarks and t-shirts. Thank you!
You can subscribe to our posts via email, email digest, browser notifications, Twitter, RSS, etc. You can filter by tag (e.g. Giveaway), keyword, author. We won't give your email address to anyone. Subscribe.

STUART STAROSTA, on our staff from March 2015 to November 2018, is a lifelong SFF reader who makes his living reviewing English translations of Japanese equity research. Despite growing up in beautiful Hawaii, he spent most of his time reading as many SFF books as possible. After getting an MA in Japanese-English translation in Monterey, CA, he lived in Tokyo, Japan for about 15 years before moving to London in 2017 with his wife, daughter, and dog named Lani. Stuart's reading goal is to read as many classic SF novels and Hugo/Nebula winners as possible, David Pringle's 100 Best SF and 100 Best Fantasy Novels, along with newer books & series that are too highly-praised to be ignored. His favorite authors include Philip K Dick, China Mieville, Iain M. Banks, N.K. Jemisin, J.G. Ballard, Lucius Shepard, Neal Stephenson, Kurt Vonnegut, George R.R. Martin, Neil Gaiman, Robert Silverberg, Roger Zelazny, Ursula K. LeGuin, Guy Gavriel Kay, Arthur C. Clarke, H.G. Wells, Olaf Stapledon, J.R.R. Tolkien, Mervyn Peake, etc.

View all posts by


  1. I think awards are a very good way to look at what was happening at a certain point in time. I can see how this would have been completely gosh-wow in 1995, with “daring” ideas like Jungian shadow-self playing out in the “cyberverse.” Racy! Almost as good as Dr Jekyll and Mr. Hyde.

    Seriously, though, the awards seem like more of a snapshot than the true test of a lasting work. I wonder if the accessibility of this — since it could easily be marketed as a thriller to more mainstream readers — nudged it towards an award as well.

  2. Perhaps I’m expecting too much of awards, and they do reflect the shifting tastes of the times, but my feeling has always been that major awards should highlight books of merit that might not otherwise get as much attention. Books by popular authors with an established fan base really don’t need the extra press. But then I wouldn’t want it to be a snob-fest either. I guess there’s really no way to please everyone, since all of our reading tastes vary so much.

Review this book and/or Leave a comment:

Your email address will not be published. Required fields are marked *