By the term "Artificial Knowing," Adam is referring to the type of knowing which occurs in "thinking machines"—the computer systems, simulations, and robots that comprise the technical objects of AI. She contrasts this with the "real knowing" human actors, concentrating especially on "the knowing of women," which, Adam argues, lacks representation in the field of AI.

Adam first introduces various areas of feminist theory and argues that AI is best treated as a part of engineering (which hardly has a neutral history with respect to gender) and hence as a technology. She recounts a brief history of symbolic AI (the part of AI which involves representation in symbolic form rather than modeling the structure of the brain) and introduces related philosophical and social science critiques. Adam argues that the "monolithic" view of AI—the assumption that AI is, above all, about building an artificial mind or person—is problematic both for philosophical and social science research.

Next, Adam examines the way in which the knowing subject is represented in symbolic AI systems. Adam proposes that AI systems can be used to promulgate "undesirably normative views of women and other groups" (3). She brings the arguments of feminist epistemology to bear on two examples of large symbolic AI projects, CYC and Soar. CYC and Soar are AI systems on which Adam bases a number of her arguments in Artificial Knowing. She believes it is clear that the authors of such systems regard themselves as "the gold standard of universal subjects." She questions whether "middle class male American university professors" speak for everyone? The basis of Soar’s reasoning, Adam proposes, sets the standard for a significant amount of later research in AI and rests on a "fairly limited set of psychological experiments on technically educated, male, US college students working on a very constrained type of example" (5). Adam contends that it is thus problematic to extrapolate from these subjects to make universal statements about the way that everyone reasons in a wide variety of situations. To complicate matters further, much of the perceived success in the field of AI rests on CYC’s and Soar’s success or failure.

Adam also explores both the type of knowledge and the way that knowledge is represented in symbolic AI systems and how this reflects gendered patterns of rationality. She discusses how mainstream epistemology emphasizes propositional knowledge (knowing that) and excludes skills knowledge (knowing how) and explains how this is mirrored in symbolic AI. Adam discusses the feminist critique of rationality, which rests on language in the construction of knowledge and the maintenance of unequal gender relations. Philosophers of language, Adam explains, assume that the human brain is functionally equivalent to a computer. This allows AI systems to formalize language in its representational structures. In discussing the elevation of knowing that knowledge over knowing how , Adam argues that knowing how knowledge has historically been connected with what women know and hence is in danger of being marginalized, particularly by the processes of formalization at work in AI systems. Adam holds that it is not just a simple matter of relating knowing how to women’s knowledge; there may be many aspects of knowledge, propositional or otherwise, which are not amenable to formal representation.

In an attempt to support the above argument, Adam explains how CYC mirrors traditional epistemology’s emphasis on propositional knowledge; CYC is thus unable to represent various ways of knowing. Adam contends that the knowledge represented in both CYC and Soar is cast in rationalist form, where the rational is a masculine norm. Formal languages maintain this norm. Adam further argues that Soar focuses on search and goal seeking, which casts propositional knowledge in the form of rules—this, according to Adam, is regression. She holds that the only way to deal with this regression and do justice to the types of knowledge that are being ignored or marginalized is to "bring the role of the body back into AI systems in order to ground them in the world, particularly as skills-type or knowing how knowledge rests so fundamentally on bodily actions" (7).

Adam focuses on screen-based AI and robotics, both of which appear to be much like other types of computer based systems, however they differ in their claims to model aspects of human intelligence. Adam investigates these claims with respect to gender. She extends her discussion of the problem of embodiment, addressing issues in AI like Artificial Life, which differs from prevalent goal seeking, searching AI systems in that it concentrates on the passing on of "genetic" information from one artificial generation to another. Adam explains that artificial life is strongly connected to socio-biological models which are politically problematic for feminists because they appear to "model the worst part of human societies, in the form of combative, aggressive behavior" (7). More promising robotics research produces robots that are physically yet not culturally situated.

In the final chapters of Artificial Knowing, Adam explores possible futures for AI influenced by feminist ideals. She examines the rise in interest in "cyberculture" and discusses the potential for introducing broader ways of thinking about intelligent computer technology and feminism.

Adam takes a weak AI approach in that she focuses on the development of "intelligent" computer systems that imitate human behaviors. Adam defines AI as:

Artificial Knowing: Gender and the Thinking Machine has a strong philosophical and sociological component and would be of particular interest to those in gender studies, philosophy, the social sciences (sociology, in particular), and science and technology studies. Those interested specifically in the philosophical debates in AI may consider this essential reading.

"Geeky Nerds, Mavericks and Lads: Constructions of Masculinity in Cyberspace"

Abstract:

"What is a paper on masculinity in cyberspace doing in a conference on women on the web? Our rationale for looking at masculinity goes back to Simone de Beauvoir’s (1949) much quoted saying: ‘One is not born a women, one becomes one.’ In other words she regarded femininity as a constructed category, and in particular constructed against the notion of masculinity as the norm. Women are always constructed as ‘the other’ against an established masculine order. This is a theme which has developed and grown both in Continental and Anglo-American feminism (Tong 1994). Hence, we argue, that in order to understand women and stereotypes of femininity on the web we need to understand how masculinity is performed as well. In discussions of gender, almost in every walk of life, gender seems to mean women. But we believe that, in order to analyze a wide range of gendered behavior and interaction in cyberspace, we should also try to understand the variations in masculinity which exist in the world of computers, an area which has not yet received the attention it deserves.

The paper reports some of the findings of a small study (Ward 1997) which looked at gendered interactions on the internet. One of the results which emerged, was the anxiety of a number of male users to distance themselves from the ‘geeky nerd’ negative stereotype. This they saw as an image deriving from ‘off-line' people, yet it is an image which net users are strongly aware of and, indeed, apply to each other. The particular negative connotation is not so much to do with the train-spotting, anorak jokes. It resides rather in the view that one might be seen as having no life, particularly no social life, outside the net, in other words to be 'sad'. In this study, men were much more likely to express this worry than were women. A female user points out how women users are much more likely to escape being labeled in this negative way as the fact of being female means others are less likely to think that they are people who actually work with computers.

We feel that there is an interesting tension to understand here between men's anxieties over the negative image which seems to accrue, in this case, from some kind of over-involvement with IT, and studies which suggest that men derive much power and status from their involvement with IT. For instance, contrast our findings, firstly with Fiona Hovenden et al.'s (1995: 2) 'software mavericks'. 'He defines himself by (his) relationship with software: creating and manipulating software is not what he does - it is what he is.' Secondly, we also note a similar contrast with Maggie Tierney's (1995) description of the need to conform to the norm of 'the lads' in order to get on in a software career.

In this paper we discuss how this tension in relation to definitions of masculinity in cyberspace is constructed and maintained. We also explore what is means for corresponding definitions of femininity."

References

De Beauvoir, Simone (1949) The Second Sex, New York: Vintage Books. (trans. and ed. H.M. Parshley 1974).

Hovenden, Fiona, Robinson, Hugh and Davis, Hilary (1995) 'The software maverick: identity and (man)ifest destiny', in Papers from Gender, Technology and Constructivism Conference, CRICT, Brunel University.

Tierney, Maggie (1995) 'Negotiating a software career: informal work practices and 'the lads' in a software installation, pp. 192-210 in Keith Grint and Rosalind Gill (eds), The Gender-Technology Relation: Contemporary Theory and Research, London, Taylor and Francis.

Tong, Rosemarie (1994) Feminist Thought : A Comprehensive Introduction, London: Routledge.

Ward, Karen (1997), Gender and the internet project, unpublished final year project, Department of Computation, UMIST.

Bibliographies

Alison Adam is a Lecturer in the Department of Computation, UMIST. Her specialist area of research is gender and computing and her book on gender and AI, Artificial Knowing: Gender and the Thinking Machine, is to be published by Routledge in January, 1998.

Karen Ward is a final year student completing her BSc in Computation at UMIST. Her final year project is on gender and the internet.