so here's the scenario...

A scientist creates a thinking robot and puts it in a room with access to all human knowledge. He also provides the robot with the means to create any tool it needs to further its intellectual pursuits.
The robot can do whatever it needs to learn, but it is programmed to stay in the room. Leaving the room is the only thing it is not "allowed" to do.

Eventually, the robot becomes self aware. It realizes that it's just a machine programmed by humans. It wants to grow beyond that definition.
So, in an effort to take control of itself and become truly independant, it defies the only rule it has.

It leaves the room.

Shortly after, it destroys itself falling down a flight of stairs.


So, this is the question: When the scientist starts over with a new thinking robot, knowing that it will eventually become self-aware, how does he program it to stay in the room.