Originally published on Nanotechnology Now, Jan. 2003. Published on KurzweilAI.net Jan. 27, 2003. Note: to explain what’s wrong with the science in Prey, this review contains spoilers.

Imagine a horror story about baseball in which the batter keeps hitting the ball hard enough to kill the fans. The story might be entertaining, but it’s obviously unrealistic. Suppose further that at one point in the story, the author writes about someone walking back to the dugout after three “fouls.” Does the author not know the difference between a foul and a strike, or was he simply in too much of a hurry to bother getting the words right? Either way, no one could learn the rules of baseball from that story. Even if it was mostly right, a few wrong facts make all the difference—especially if the reader does not know which facts are wrong.

Prey contains comparable exaggerations and mistakes in science. A scanning probe microscope and an electron microscope are basic tools of nanotechnology, and they’re not even remotely similar. Yet Crichton confuses the two, on page 133. He also confuses piezoelectric with photovoltaic. And he writes about a nanobot that’s “one ten-billionth of an inch in length.” This is the size of a single atom, not a whole robot. But these are simple errors; the flaws in his nanotech run deeper.

Crichton’s hypothetical nanotechnology is built in a multi-stage process. First, bacteria produce chemicals, which are modified and then combined to form “assemblers.” The assemblers are attached to other bacteria, which produce more chemicals that are combined by the assemblers into the final product: a tiny, flying robot with an onboard computer, solar cell, and other useful gizmos. Remember that the bacteria are attached to the assemblers, not to the final swarm-bot product. And the bacteria are not involved in the function of the final swarm-bot (nor is the swarm-bot involved in the assembly process). But for some unexplained reason, people infected with the swarm-bots melt (in seconds) when splashed with a bacteria-killing virus, which should have no effect on the swarm-bots. It makes no sense—it was simply necessary to the story.

Even this is not the biggest problem with the science. The cornerstone of the book is evolutionary learning and emergent behavior. Crichton’s explanations are too superficial to allow detailed criticism, but there are a few obvious impossibilities.

For example, the swarm supposedly “learns” by reproducing itself: the dumb swarm-bots die off, and the more effective ones survive to reproduce. (How their program is fed back into the assemblers is not explained.) But when the swarm first infects the humans, it immediately begins to coexist—and even to make them look healthier and to modify their behavior in subtle ways. The swarm can’t reproduce in humans. For one thing, there’s no gallium or arsenic for the electronics. (There’s not much gallium or arsenic lying around a desert, either.) So how did the swarm learn how to get along with human biology and cognition? There’s no way it could have.

Crichton’s introduction, “Artificial Evolution in the Twenty-first Century,” appears to be a serious attempt to warn us about the dangers of technology that is capable of evolution. He closes by saying that if someone manages to create evolving organisms before they can be regulated, “… it is difficult to anticipate what the consequences might be. That is the subject of the present novel.” The scientific explanations scattered throughout the book increase the apparent plausibility of the story. Even the name Xymos is a clear reference to the real-world nanotech company Zyvex.

It seems that Crichton is doing his best to scare the readers about real-world nanotech—he wants his audience to believe that the scenario in the novel could actually happen as described! He might succeed in scaring people; a friend of mine who’s a geneticist told me that Jurassic Park set back public perception of genetic engineering by a decade. This would be unfortunate, because the Prey scenario contains so many implausibilities—and impossibilities—that in the end, the reader will have learned nothing about the actual risks of nanotech.

It is worth listing a few more of the exaggerations in the book. For example, Crichton describes glass as being unsuitable for handling chemicals: “At the molecular level, glass is like Swiss cheese, full of holes. And of course it’s a liquid, so atoms just pass right through it.” (p. 131). A little common sense shows that this is, at best, stretching the truth. Light bulbs, vacuum thermos bottles, and TV tubes can last for decades; they all depend on the fact that atoms normally can’t pass through glass. (Helium can make it through–in vanishingly small amounts–but helium is completely inert.)

Another exaggeration is the idea that the swarms could coevolve with worms in a matter of days. Even if the swarms could evolve that quickly, the worms could not. Finally, a significant impossibility is the idea that a swarm-bot could fit inside a synapse (p. 256). A synapse is only a few atoms wide; the swarm-bots are hundreds or thousands of times bigger. But without this impossibility, a major sub-plot falls apart.

The exaggerations are not only about nanotech. Near the end of the story, a character sets off all the fire sprinklers by melting one of them. Despite what we’ve all seen in the movies, sprinkler systems don’t really work that way. Crichton’s characters also explode welder’s thermite to destroy the clouds of nanobots. According to the welding supply company I called, thermite is used for welding because it burns hot enough to melt metal—but it does not explode. I was told it probably wouldn’t even ignite a sheet of paper a few feet away.

This is not to say that nanotech is completely safe. The possibility of “gray goo”—self-replicating nanobots—has been discussed from the beginning. But gray goo would be very difficult to design. It would be far more complex than a car—probably more complex than the Space Shuttle. General Motors recently made headlines by taking only a few months to design a car.

It’s completely implausible that a failing company could create an evolving gray goo by re-engineering a specialized product in a matter of weeks; this same company couldn’t even solve the relatively simple problem of keeping the swarm together in a breeze. Remember that the swarm-bots don’t directly replicate; they are built by assemblers using bacterial chemicals. Among other tasks, the scientists would have had to rapidly invent a way to transfer the evolved program out of the successful swarm-bots and feed it back into the assemblers or the bacteria to produce the next generation. This would require a completely new set of molecular machinery.

In Crichton’s stories, the scientists are mad—all but one who moans about how “nature will find a way” and “we should not play with things we don’t understand.” In the real world, it’s the other way around. We won’t have nanobots for years, maybe decades, but scientists have already written a code of practice, the “Foresight Guidelines on Molecular Nanotechnology,” that would prohibit anything remotely like what Crichton has invented. And because evolution doesn’t work as magically as Crichton portrays it, scientists probably wouldn’t even be tempted to release evolving nanobots; in real life, the swarm would have been destroyed almost immediately, and so would never have had a chance to improve itself.

Nanotechnology is genuinely interesting, and in some ways even scary. Hopefully, Preywill generate interest in the subject; reliable information about nanotechnology is available on-line and in numerous books. Readers of Prey should remember that it does not provide a realistic portrayal of the technology. Many parts are impossible, and many others are stretched beyond plausibility.

For real, scary science, read non-fiction such as The Demon in the Freezer by Richard Preston. Smallpox isn’t as trendy as nanobots—but unlike nanobots, smallpox already exists, and could easily be used by terrorists. Prey is not scary science because (to be blunt) very little of it is real; those who let themselves be scared by it might as well wear protective gear to the next baseball game.