This is absolutely fascinating:
Researchers at the Ecole Polytechnique Fédérale de Lausanne in Switzerland have found that robots equipped with artificial neural networks and programmed to find “food” eventually learned to conceal their visual signals from other robots to keep the food for themselves. The results are detailed in an upcoming PNAS study.
Although I have to wonder, is it really a lie or is it just evolved selfishness as a mechanism for “survival”? The robots are programmed to find food, a kind of “find the food or starve to death” directive. I don’t find it surprising that these robots determine they have a better chance of survival if they don’t tell everyone else where the food is. I wonder if some will “evolve” to signal where poison is, in order to bump off the opposition. If this becomes the case, it starts raising interesting ethical questions about neural networks.
Maybe all those films where the neural network decides humankind is a threat or unnecessary to its own survival aren’t too far off the mark…