Geoff's Laws of Robotics: Corrolaries

  • Corollaries to the Geoff's first law of robotics:

    1. It doesn't care what you thought you said.

    2. The way in which the robot interprets instructions is up to the robot, not up to you.


  • Corollaries to the Geoff's second law of robotics:

    1. Robots don't understand English.

    2. If robots simulate understanding English, the English they understand is not the same as the English you speak.

    3. If the language seems to be clear and straightforward to you, this is an illusion.

    4. No matter how well-documented the language, some commands are always undocumented.

    5. The undocumented commands usually include "halt and catch fire."


  • Corollaries to the Geoff's third law of robotics:

    1. Causing harm to a human, or through inaction causing a human to come to harm, is of no particular concern to the robot.

    2. It doesn't care about whether it causes harm to itself, either.

    3. Those "keep out" zones are there for a reason.

    4. Artificial intelligence isn't.




  • page by Geoffrey A. Landis, 2004