A demonstration in PAGI World of an agent solving a simple moral dilemma. For more information, see: rair.cogsci.rpi... Video: Larry Li, John Licato, Selmer Bringsjord Rensselaer AI and Reasoning (RAIR) Lab
In the last example, is the problem really resolved? Now that it splits up the resources, it still has to choose between delivering the resource to either the left or right individual first. I guess if the conditions for the left and right individuals were the exact same, it would just choose at random or perform some kind of multi-delivering method to ensure both receive their treatment as fast & efficient as possible. Regardless, thanks for sharing this demonstration! :]
LucidEnigma21 You're exactly right. If there is no criteria it can use to choose one over the other, it still has to choose one, probably randomly. At the very least, it now "knows" that it's in such a dilemma state and will have to accept that no matter what it does, one of its obligations will be broken. That seems better than freezing up and breaking both obligations.
