Technology is moving very fast. Giant tech companies are building smart machines. These used to seem like science fiction. But as robots and artificial intelligence (AI) get smarter, they raise a big question: Who is blamed if a robot commits a crime?

If you tell a robot to break the law, does the robot go to jail? Or do you? This guide breaks down how criminal law looks at robots, who is truly blamed, and why human intent is still the key to justice.

The “Innocent Instrument” Rule in Criminal Law

To understand robot crime, we must see how the law defines a “criminal.” A robot is a machine made of metal, plastic, and code. It is not a human being. Because of this, a robot cannot be arrested, put on trial, or locked in a prison cell. But this does not mean you can use a robot as a shield to commit crimes.

Blaming the Person Behind the Controls

In criminal law, there is an old rule. It is called the “innocent agent” rule. This rule applies when a person uses a child, an animal, or a tool to commit a crime.

For example, if a person trains a dog to attack someone, the dog is not put on trial. Instead, the person who trained and ordered the dog faces criminal charges.

The law views a robot exactly the same way. The robot is just a tool. If you tell your smart robot to enter a store and steal an item, the law treats it as if you stole it yourself. Prosecutors will focus on the human who gave the orders, controlled the machine, and had the intent to steal.

What Happens if a Robot Acts on Its Own?

The legal questions get much harder with true AI. Smart systems can learn and adapt on their own. What happens when a robot does something its owner never told it to do?

If an autonomous system makes its own choice to break the law, finding the guilty party is a major challenge.

The Problem of Autonomous Decision-Making

Usually, to convict someone of a crime, prosecutors must prove two things:

1. Actus Reus: A physical wrongful act was committed.
2. Mens Rea: The person had a guilty mind or criminal intent.

If an AI system makes its own choice to commit a crime, the human owner might not have any criminal intent. But the owner or programmer could still face charges under another legal rule. This is called the rule of “natural probable consequences.”

This means that if you use or program an AI in a highly reckless way, and you should have known it could cause harm, you can be held responsible. For example, if you let an untested delivery robot roam a crowded sidewalk and it hurts someone, you could face reckless endangerment charges.

Can You Get a DUI While Operating a Robot?

Another interesting question is whether you can get a DUI while operating a robot while drunk.

In Pennsylvania, the law says you can get a DUI if you are driving or in “actual physical control” of a vehicle. Simply telling a small kitchen robot to take out the trash while you are drunk is not a DUI.

But what if the robot is large and designed to carry people or heavy cargo? As smart machines grow and look more like traditional vehicles, courts will have to draw a clear line. If you are riding or controlling a large robotic transport machine on a public road while drunk, you could face a DUI.

In the end, technology is not a shield from criminal charges. If you use a machine to cause harm, the law will look past the computer to find the human controller.

This article is for educational and hypothetical discussion only and does not constitute legal advice. Tesla and Google are referenced solely for purposes of discussing developments in technology and artificial intelligence. Neither company is affiliated with or has endorsed Attorney Michael Kotik or this article.