The social implications of artificial intelligence.

Abstract

For 18 years. I have been publishing books and papers on the subject of the social implications of Artificial Intelligence (AI). This is an area which is has been, and
\nremains, in need of more academic attention of a serious nature than it currently receives. It will be useful to attempt a working definition of the field of AI at this stage. There is a considerable amount of disagreement as to what does and does not constitute AI and this often has important consequences for discussions of the social implications of the field.
\nIn brief, I define AI as the study of intelligent behaviour (in humans, animals, and machines) and the attempt to find ways in which such behaviour could be engineered in any type of artefact. My position on the definition of AI as a field of activity is set out in full in various places in the works submitted for this application. Most important are
\nChapter 2 of Whitby, 1988b, Chapter 3 of Whitby, 1996, and Whitby, 2000. This definition is distinctive (though not unique). For the purposes of discussion of social
\nimplications, its most distinctive feature is that it does not require the imitation or replication of human intellectual attributes. Because, under this definition, AI is not limited to the study of and attempt to build human-like intelligence the discussion of its social implications is rendered much broader.
\nAlso, because AÏ encompasses the attempt to engineer intelligent behaviour in any type of artefact, discussion of its social implications will need to consider the way in which AI technology, methods, and attitudes can permeate other different areas. This will include a wide range of technologies which include an AI element and a wide range of disciplines which are influenced by AI ideas. Thus the social implications of AI are turned into an immensely important field of study, since AI technology will steadily continue to permeate other technologies and thereby society as a whole. Many of the social implications of this technological process are nonobvious and surprising. If we are to make sensible, timely, and practical policy decisions
\nand legislation then it is important to be as clear as possible about likely technological developments and their social implications.
\nWe may initially attempt to characterise various approaches by other authorities on the social implications of AI. Thèse range from the wildly spéculative such as Warwick
\n(1988) and Moravec (1988) to the mainly technical, for example Michie (1986). At the wildly spéculative end of this continuum represented by Professor Warwick there are
\nscare stories involving robots taking over the earth. (See for example Warwick, 1998 pp. 21-38) At the other end of the continuum, there are writers who often see AI as entirely positive, or as having no social implications at all.
\nMost authorities will, or at least should, occupy a position somewhere between these extremes. However, in giving serious académie considération to this area, one needs to respond to this entire range of approaches. That is to say that one must (as a minimum) both be conversant with probable technical developments and also carefully and critically respond to speculations about the nature of future society. In my research I have consistently attempted to do just this.
\nThis is obviously a cross-disciplinary exercise and the differing methodologies of different disciplines présent further problems in determining the best (or an
\napproximation to the best) approach. For a number of reasons, which will be fully explored in this statement, my approach has concentrated (mainly, though not exclusively) on the attempt to provide guidance to those actually concerned with the technical and scientific development of AI.
\nThe published books and papers submitted as part of this application span a period of 16 years. Thèse works form a cohérent body of research around the area of the social
\nimpl