ABSTRACT
Public interest and support for strong AI has lagged because optimistic predictions often failed and because progress is difficult to measure or demonstrate. This chapter argues, however, that the next stage of development of AI, for at least the next decade and more likely for the next 25 years, will be increasingly dependent on contributions from strong AI. This hypothesis arises from an empirical study of the history of AI in several practical domains using what Abbott (2004) calls the small-N comparative method. This method examines a small number of varied cases in moderate detail, drawing on descriptive case study accounts. The cases are selected to illustrate different processes, risks, and benefits. This method draws on the qualitative insights of specialists who have studied each of the cases in depth and over historically significant periods of time. The small-N comparisons help to identify general patterns and to suggest priorities for further research. Readers who want more depth on each case are encouraged to pursue links to the original case research.
