Asimov's Three Laws

You’ve probably heard some version of this. “Don’t worry, we’ll just give the robots rules.”

Cool. And while we’re at it, let’s give toddlers a policy memo and see how that goes.

The setup

“Asimov’s Three Laws” comes from sci-fi writer Isaac Asimov. He imagined robots hardwired with three basic rules:

• Don’t harm humans.
• Obey humans.
• Protect yourself.

In that order. So if obeying a human would hurt someone, the robot is supposed to ignore the command. Nice. Neat. Sounds like robot seatbelts.

It’s one of those ideas that is so elegant people keep dragging it into real AI conversations like it’s a practical safety plan, not a fictional plot device.

The problem

The laws sound simple until you ask one annoying follow-up question. What counts as harm?

If a robot takes your car keys because you’re drunk, that helps you and ruins your night. If it tells a brutal truth, that might help long term and hurt right now. If it does nothing while you make a dumb decision, is that safer or worse?

This is the IKEA-furniture problem of ethics. The picture on the box looks clean. Then you open it and suddenly there are 97 screws and one weird wooden peg nobody trusts.

Humans don’t even agree on these tradeoffs. We argue about parenting, medicine, driving, war, office Slack messages. And now you want a machine to sort it out perfectly?

Why people still mention it

Because it’s a great shorthand. “Build AI that won’t hurt us” is a real goal. Asimov gave that anxiety a catchy wrapper.

But the Humanoids glossary version nails it. These are fictional robot ethics rules, endlessly cited and entirely unimplementable.

Why it matters

If someone says AI safety is solved with “just add rules,” you should hear alarm bells. Ethics isn’t a light switch. It’s more like cooking for picky relatives with food allergies. One wrong move, and everybody’s upset.