What will humans be like generations from now in a world transformed by artificial intelligence (AI)? Plenty of thinkers have addressed this question, considering how AI will alter lives—often for better, sometimes for worse.
They have conjured dramatic scenarios, like the AI-driven extinction of humans (and many other species) or our assimilation into human-AI cyborgs. The predictions are generally grim, pitting the fate of all humans against a unitary (or unified) AI opponent.
What if the AI future doesn’t stretch to these sci-fi dystopias? For an evolutionary biologist, seeing AI technologies diversify into all applications is like seeing microbes, plants, and animals multiply in an ecological landscape.
This led me to ask: How might human evolution be altered by interactions with a world of rich AI diversity? In a paper just published in The Quarterly Review of Biology, I considered the many ways AI might alter physical, biological, and social environments and how that might influence natural selection.
Natural selection—the mechanism behind evolution—is an inevitable consequence of genetic differences in reproduction among individuals.
Those differences arise as a result of interactions with physical features of the environment (like minimum temperatures), with other species (like predators or parasites), and with other members of the same species (like mates, allies, or hostile outsiders).
When Asian gray wolves started hanging around humans around 30,000 years ago, the more reactive wolves were chased away or killed. This whittled away genes for skittishness and aggression, beginning the process of dog domestication.
The inadvertent selection that turned wolves into dogs is instructive in how AI might inadvertently shape the evolution of human brains and behavior.
“Trying to predict the future is a mug’s game,” said English author Douglas Adams. This is especially true of technologies like AI.
But predicting evolution is, if anything, even more precarious. Combining the two involves considerable speculation and the very strong possibility of being wrong.
At the risk of being wrong, my intention is to start a conversation about how AI might alter human evolution and the traits that we most value in one another.
