Limits of ‘expert systems’ for political-economic forecasting

There is an apparently growing belief that knowledge can be codified into ‘expert systems’ that can systematically convert sociopolitical information into probabilistic statements about the political and economic future. The development of expert systems in medical diagnosis, mineral exploration, aircraft carrier control systems, and so on, has fueled this optimism. In a sense, the question is whether forecasing can, will, or should be handed over to AI. Will the issues of Technological Forecasting and Social Change 20 years from now be filled with computer code like the pages of PC World? The logic of how human experts apply judgment and how these automated systems represent ‘knowledge’ lead to a resounding endorsement of keeping the humans employed. But the challenges of automated systems open up fascinating challenges for improving the management of human experts.