Could a robot become president?

A bot certainly wouldn’t need breaks for sleep or snacks—but it also might not care much about the people it’s leading.

Artificial intelligence (also called AI) is beginning to take on all kinds of jobs that were once only for humans. Computer systems can diagnose diseases, research legal cases, and even write songs. AI can do many of the tasks people can—and sometimes do them better.

So is it possible that someday, we might have a robot do one of the most important jobs of all: leading a nation?

Mind the machine

The idea might sound far-fetched. But in 2016, one group pushed for Watson, the Jeopardy!-winning AI developed by IBM, to run for president of the United States. (Jeopary! is a trivia-based game show.) They thought a bot could make a good leader. Humans can be influenced by anger, prejudice, and ego, but robots would—in theory—use only the facts when making a decision and be influenced by their emotions.

People who support voting for a robot also argue that computers are better than humans at processing huge amounts of information. A human president can’t keep up with every detail coming in from government agencies, the economy, the people, the press, and more. But artificial intelligence can.

For instance, to answer each Jeopardy! question, Watson instantly analyzed 200 million pages of information—including all of Wikipedia. A future robot president could go further, basing every decision on all of human knowledge. Plus, a robot would work harder than any human in history, never stopping to sleep or get a snack.

Brainy bot?

Even most people in favor of a robo-president say our current AI isn’t ready yet. Our most advanced programs, such as Watson, can only do one task. A robot ruler would need to solve many kinds of different problems, from how to negotiate with other countries to how to handle the changing climate.

Another hurdle is that computers aren’t great communicators. They can process massive amounts of information to make a decision—but they can’t explain their thought process afterward. So a human president can talk through their actions to the people, but a computer might leave the country confused.

As for prejudice, AI is currently only as unbiased as its programming. For instance, AI programs have shown trouble distinguishing between light-skinned and dark-skinned people, or people of different genders.

Some experts say AI will never measure up. Others think it will overcome these obstacles within the next few decades.

Would you vote for a robot?

Excerpted from Nat Geo Kids’ Ultimate Book of the Future by Stephanie Warren Drimmer