Thought experiment: Asimov’s First Law of Robotics

Asimov’s first law of robotics states, “A robot may not injure a human being or, through inaction, allow a human being to come to harm.” This sounds reasonable, so here’s a thought experiment about designing some robots to implement it.

The first robot has a camera and a water pistol, together with the ability to move around. This robot uses the camera to detect when people are smoking and uses the water pistol to put out the cigarette (or whatever the person is smoking). This robot is taking positive action to avoid people harming themselves.

The second robot has a similar design, except the water pistol is filled with a combination of cod liver oil and cayenne pepper. It uses its camera to detect when someone is eating something unhealthy, then it shoots the food with the cod liver oil / cayenne mixture. Again, the robot is taking positive action to prevent people from harming themselves.

These two robots are simple (technically very easy to build), not intelligent (no AI required), follow Asimov’s first law, and yet they are socially unacceptable. Any actual intelligent robot would need to take actions like these and even more intrusive actions to prevent people from harming themselves.

Maybe we need something more like “A robot shouldn’t actively harm people, nor should it interfere with their choices.”

Leave a Reply

You must be logged in to post a comment.