Markup View

Author: randomsurfer

Jan 15th 2013 at 2:24:02 PM

This is a concern of Susan Calvin in one of IsaacAsimov's short stories in ''Literature/IRobot''.  A robot has been built with a modified First Law, which in its case permits a robot to allow a human to come to harm via inaction.  Calvin posits a situation where a robot with this modification can commit murder - by dropping a heavy weight above a human, knowing that its quick reflexes will allow it to catch the weight in time to not harm the human; but then having dropped the weight it has the ability to decide not to catch the weight.

[Edited because of Tiiba's comment below.  I don't know if Tiiba was specifically referring to my example, but looking at it I could see that it looked like I meant that the First Law ''does'' allow this.  It does not.  The robot in question had a modified First Law because it was working with humans who were being exposed to brief amounts of radiation.  Other robots would rush in to save the human because he might forget to leave the radiated area in time, even though the radiation is instantly lethal to a robot brain.]