The AI Apocalypse Is Coming...

But Comparative Advantage Will Save Us

Artificial Intelligence will make many human jobs obsolete. Computers are extremely fast and accurate at solving equations, data processing, and are getting much better at reading and understanding text. The routine tasks of many Lawyers, Accountants, and even programmers will soon be done by computers with AI. The scary thing is that these Artificial Intelligences are only getting smarter, and they are improving at a much faster rate than our puny biology. Eventually, there will be nothing that a human can do that computer running an AI program can’t do faster, including writing fiction, making music, and giving care. Once we pass this event horizon, human civilization will collapse into a singularity and all biological life will be cast aside as useless by our AI overlords.

Except we won’t be useless. Even when computer programs can do everything we can do quicker and more accurately, there will still be mutual gains from specialization and trade. Let’s look at this in an example. We can imagine a simplified market with AIs and Human producers who can solve mathematical equations or babysit dogs. These super-advanced AI are better than humans at everything, so the production table for one hour of work looks like this:

A human can watch at most 5 dogs in an hour or solve 20 equations. The AI is much better at both tasks. It can solve ten million equations in an hour or watch 10 dogs at once. Even though the human is worse at both tasks, resources are only maximized when the tasks are split up between the AI and human worker. The AI’s opportunity cost for watching one dog (what else it could have done with the time it spent watching the dog) is a million equations while the opportunity cost of a human worker watching one dog is only four equations. Each of the producers can specialize in what they have the lowest opportunity cost in producing and trade with the other to get more than they could have through self-sufficiency.

This simple example represents an important general principle: Absolute advantage, even very large advantage, is not enough to make producers obsolete. AI will continue to get better than humans at more things, but we need not worry. Comparative advantage will guarantee the mutual profitability of human-AI partnerships even when they are orders of magnitude better than us at all tasks.