Anything that thinks logically can be fooled by something else which thinks at least as logically as it does. The easiest way to fool a completely logical robot is to feed it the same stimulus sequence over and over again so it gets locked in a loop. This was best demonstrated by the famous Herring Sandwich experiments conducted millennia ago at MISPWOSO (the Maximegalon Institute of Slowly and Painfully Working Out the Surprisingly Obvious).
