May I be permitted to quote myself from a recent post ? :- Restrictions on A.I. development have been called for by many of those working at the frontiers of its development. But as long as there are no prohibitions, many of these same developers feel they must continue to try to get a lead on the others. They are not talking about what most people think of A.I. The sort that can lighten your work-load, by searching, finding, analysing, and summarising. Although even that can go off the rails. If the sources of information, the algorithms, etc. are not correctly chosen. E.g. an A.I. enquiry as to what sort of leader would be best for Europe returned the unequivocal answer: "HITLER". What the experts are worried about is Artificial General Intelligence (A.G.I.) and Artificial Super-Intelligence (A.S.I.). [If need be, do some homework]. We seem to be getting very close.
