I, Robot and Game Theory
Posted by Patrick on April 8, 2009
- A robot may not injure a human being or, through inaction, allow a human being to come to harm.
- A robot must obey orders given to it by human beings, except where such orders would conflict with the First Law.
- A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.
The theme of the movie was probably based mostly on the last story, called The Evitable Conflict . In this story, powerful positronic computers around the world have the task of optimizing the world’s economy. Naturally, such a complex problem cannot be solved without making trade offs and sacrificing the well being of some individuals for the greater good. As such, the machines come to generalize the First Law to mean: “No robot may injure humanity or, through inaction, allow humanity to come to harm.” The three laws, which were supposed to prevent robots from taking over, in essence, dictate them to do so (for our own good).
In the movie, the detective uncovers this scheme soon enough, and with the help of a non-three-laws robot, manages to shutdown the main computer in time. In the book, however, humans realize that the solutions the machines are implementing are the best, and any deviation from their scheme would leave us worse off overall. And so, they come to accept the benevolent (and selfless) dictatorship of the machines.
When I read this story, I wondered if a “best solution” could exist to such complex problems as “optimizing the world’s economy”. I was recently reminded of this question after watching political scientist Bruce Bueno de Mesquita speak at TED  about how game theory  can be used to predict the most likely outcome of a situation where many players are trying to optimize their own self interest.
In the language of game theory, it seems that Asimov’s machines were finding some kind of equilibria. But with all the machines cooperating not for their own self interest, but for that of humanity, would game theory be the proper tool to use?
Finally, Cory Doctorow puts a spin on things with his I, Robot , and imagines a world divided into two: In the West, robots are bound by the three laws; in the East, the are not. Maybe because of this division, his three-laws-bound robots don’t seem to generalize the first law. His non-three-laws robots, however, open up a whole new set of possibilities…
- Wikipedia: I, Robot, <http://en.wikipedia.org/wiki/I_robot>
- Wikipedia: Three Laws of Robotics, <http://en.wikipedia.org/wiki/Three_Laws_of_Robotics>
- Wikipedia: The Evitable Conflict, <http://en.wikipedia.org/wiki/The_Evitable_Conflict>
- TED: , <http://www.ted.com/index.php/talk/bruce_bueno_de_mesquita_predicts_iran_s_future.html>
- Wikipedia: Game Theory, <http://en.wikipedia.org/wiki/Game_theory>
- Craphound: Overclocked , <http://craphound.com/overclocked/download>
- Wikimedia File, <http://en.wikipedia.org/wiki/File:I_Robot_-_Runaround.jpg>