Ph.D. in Philosophy by the Pontifical Catholic University of Rio Grande do Sul
The relationship between gender and knowledge is one of the main topics of debate in the feminist movement. More specifically, the demand for fair and equal access for women to education and the possibility of producing knowledge, already given in the firsts texts considered “feminist”. In "A Declaration of the Rights of Women" (1791), Wollstonecraft asserts that "the rights for which women, along with men, must strive [...] are a natural consequence of their upbringing and their position in society". The search for breaking the imprisonment of women in the private and domestic space remained the basis of feminist thought - whether for inclusion in education and democratic processes of the first wave, for insertion in the labor market and autonomy over the second wave, or for equal opportunities of different women in particular situations to the public sphere by the third wave of the movement. Despite the countless 'glass roofs' that the movement has broken, we, women, still remain on the sidelines of various spaces of knowledge, such as in academic faculty, in senior positions in private companies, in government leadership, and even in fields that still remain seen as masculine: such as science and technology.
Parity in the participation of women, non-whites, LGBTQ+ and other marginalized groups in saber fields is, without a doubt, a requirement for us to achieve some model of a fair and independent democratic society. Still, that these groups in certain areas finds a place of resistance and discomfort due to heteronormative, misogynistic and racist structures, which alienate these subjects from these activities, are already widely claimed.
The provocation of the question “Can women be cyborgs?”, however, goes beyond the issue of equal access for women in the field of technology development and artificial intelligence - aims to research whether gender issues bring implications to the knowledge content itself, this is, how perceptions of gender distinctions affects the goal for universal content and transcendent knowledge, especially in artificial intelligence systems, that seems to make definitions of sex and gender obsolete.
Feminist Epistemology: A Critique of Disembodied Universal Reason
In “The Man of Reason” (1985), Genevieve Lloyd points to the masculine character that notions of rationality and knowledge hide behind the presuppositions of universality and objectivity throughout the history of philosophical thought. According to the author, "our belief that reason does not know any sex has, I argue, to a large extent, deceived itself", since the separation of forms of knowledge from that 'rational', from 'passionate', allowed for exclusion or hierarchization of the content of knowledge, which, due to the gender distinctions that structure society, produce “not only practical but also conceptual reasons for the conflicts that many women experience between Reason and femininity” (LLOYD, 1985, p. 10). In line with this perspective, Alison Jaggar, in “Love and knowledge: Emotion in feminist epistemology” (1989), clarifies the imprisonment of notions of knowledge in dichotomous models, which remove sensitive experience from the possibility of informing objective knowledge, contributed to the valorization of certain knowledge concepts:
“Normally, though not invariably, the rational was contrasted with the emotional, and this pair contrasted then often linked to other dichotomies. Not only has reason been contrasted with emotion, but it has also been associated with the mental, the cultural, the universal, the public and the masculine, while emotion has been associated with the irrational, the physical, the natural, the particular, the private. and, of course, the feminine" (JAGGAR, 1989, p. 151).
The field of feminist epistemology has recently focused on facing the problem of gender against the objectivity of knowledge. For many individuals who are outside the production of science, technology and knowledge, thinking about it as disinterested, impartial and without value judgment, seems to confront their experience with the history of modern science.
Sandra Harding in “Whose science? whose knowledge?: thinking from women's lives” (1986), articulates three important and distinct feminist epistemological programs:
Empiricist feminist philosophy - which seeks to correct “bad science”;
Feminist point of view - which builds knowledge from the particular experience of women;
Postmodern feminism - which suspects commitments from science and epistemologies to the Enlightenment project (HARDING, 1986, p. 11).
These epistemological programs bring, in different ways, questions and contribution to a kind of feminist method, an investigation that commits itself to a critical analysis of objectivity that reveals the masculine standards of methods of knowledge. For Harding (1986, p. 41):
“One way of looking at this problem is to realize that although scientific methods are selected, they say, precisely to eliminate all social values from inquiry, they are really operationalized to eliminate only those values that differ within whatever is considered by the community of scientists. If values and interests that can produce the most critical perspectives on science are silenced through discriminatory social actions and practices, the standard, narrowly conceived as a scientific method, will not stand a chance of maximizing value neutrality or objectivity”.
The tension exposed by the criticisms of feminist epistemology, both to social scientists and to the natural sciences, was that of the relationship between power and knowledge. Several critics focused on exposing that the universal scientific method, in fact, investigated and debated problems related only to male experience, that scientific narratives served the interests of white men, and that the entire symbolic order through which knowledge is claimed were articulated privileging the masculine and conceptualizing the feminine only as that which lacked masculinity. The criticism of objectivity, that knowledge of the world as it is, regardless of the knowing subject, sought to demonstrate that such knowledge inevitably carries traces of subjectivity in the methods taken as universal.
Many of these feminist methodologies have pointed to the situated factor of knowledge. A situated epistemological view, as several authors argue, does not represent a necessary relativism, but proposes as a methodological starting point that objectivity and subjectivity are not two oppositions in a scientific investigation, but can be constructed dialectically (see FIUMARA, LAZREG, 1994). That is, that objectivity and subjectivity are in a constant process of formation.
In “Situated knowledge: the question of science for feminism and the privilege of partial perspective” (1995), Donna Haraway argues that “the alternative to relativism is partial, locatable, critical knowledge, supported by the possibility of networks of connections, called solidarity in politics and shared conversations in epistemology” (HARAWAY, 1995, p. 24), thus, situated knowledge is not in a fixed place of opposition to that which is totalitarian, universal; but, in fact, it is in a "practice of objectivity that favors contestation, deconstruction, network connections and the hope in the transformation of knowledge systems", which can know the world in a less organized way based on axes of domination of knowledge. Haraway's proposal is for a different view of objectivity, and here I say 'vision' because of the focus given by the author to the corporeal and sensory factor of vision to place us in the world, not in a strictly passive way towards the object of knowledge, but as “active systems of perceptions, building translations and specific ways of seeing” (HARAWAY, 1995, p. 22). Vision, emphasizes Haraway, depends on what we can see, on systems of perceptions that produce meanings in the world, on positioning itself as an eye that sees, translates, describes and reproduces. With that, what one wants to discard is the totalitarian production of knowledge, the one that disguises its masculine position, in the universal and in its entirety, to the detriment of a vision of the:
“[...] knowing self [which] is partial in all its forms, never finished, complete, given or original; it is always imperfectly constructed and stitched together, and therefore capable of joining with another, of seeing together without pretending to be another. Here is the promise of objectivity: a scientific expert does not seek the position of identity with the object, but of objectivity, that is, of partial connection” (HARAWAY, 1995, p. 26).
A critique of objectivity, however, does not imply that we put the baby out with the bathwater - as Marnia Lazreg well points out in "Women's experience and feminist epistemology" (1994), a question that remains interesting for the critique of disembodied science, that does not allow an experience to be considered a form of knowledge, is whether all experience is valid in the same way as knowledge, that is, whether a 'feminist scientific method' contains or needs to contain well-defined scientific criteria for analysis. As Haraway rightly puts it, relativism falls into the same general problem as objectivism, a “trick of God”: while objectivism is knowledge of nowhere (transcendent), relativism promises to be knowledge of all places at the same time and equally (omnipresent). Knowledge needs to be seen as partial and situated, not relative, in a process of emergency.
After all, an epistemological view of women's experience must also not fall into the essentialist traps that feminism seeks to criticize, that is, it cannot be taken as a fixed and total situated place, since, as we have seen in recent debates, especially after the In the third wave of the feminist movement, the category 'woman' as a specific position of knowledge ignores the multiplicity and variety of experiences and intersections of human experience, which permeate issues beyond gender, such as race, ethnicity, nationality and sexuality, for example. Lazreg's conclusion highlights that the:
“The choice is not between science and experience, objectivity and subjectivity. The point is to realize that objectivity is an ever-increasing goal and striving to achieve it, is a never-ending historical process” (LAZREG, 1994, p. 59).
Scientific investigation committed to feminist criticism seems to point to the insufficiency of the majority models that establish criteria of rationality and objectivity as a vision situated nowhere, but capable of seeing the research object in a total way. This claim to universality implied in the most diverse fields, the social, natural and exact sciences, is obscuring a pattern that favors a male view of the world. With this, if we have a recent recognition that knowledge is situated in a world structured by gender configurations, after all, scientists who see, translate and represent the reality of the world are situated in these gendered places.
An intersection point between gender, robotics and artificial intelligence for the debate on knowledge in science and technology emerges as these systems seem to be free from gender boundaries and, therefore, from their consequences in the result of forms of knowledge. Furthermore, can we consider in what sense these technologies tend to influence our current concepts of sexual and gender differences?
Genres and Artificial Intelligence
The description made by Lloyd in "The man of reason" (1985) points to the search, which permeates the history of philosophy, for a kind of knowledge and pure reason, free from any influence of human passions and corporeality. Somewhat, we can see artificial intelligence systems as the way in which this 'pure reason' reaches its greatest potential. The question that arises as we commit ourselves to a feminist methodology of epistemology as described is, after all, can there be gender biases in the field of Artificial Intelligence?
At first, the narrative of these disembodied systems that cannot be classified as female/male, masculine/feminine, man/woman, seems to bring hope for a kind of knowledge that goes beyond the limits of these Enlightenment dichotomies.
Certainly, as claimed by women in the fields of science, technology and the development of Artificial Intelligence, the lack of parity in the participation of women, as well as other marginalized social categories, imply that many of the developed systems seek to respond to strictly male demands, alienating the problems of other subjects from the technological field. This does not imply, at least not necessarily, that Artificial Intelligence, as well as that of objectivity that Lloyd called 'man's reason', can hide in the pretension of neutrality a masculine knowledge.
Techno-feminist debates starting in the 1990s, initiated by Judy Wajcman, emphasize that “male-dominated technologies conspire to diminish the relevance of 'women's' technologies such as horticulture, cooking and childcare”. Ferrando in “Is the post-human a post-woman? Cyborgs, robots, artificial intelligence and the futures of gender: a case study” (2014) analyzes that the networks of historical connections that contextualize and situate the conception of the 'human' also form a history and situate the understanding of what constitutes these technological systems, like that of a cyborg, for example.
Also, Susan Leavy, in “Gender Bias in Artificial Intelligence: The Need for Diversity and Gender Theory in Machine Learning” (2018), argued that gender biases can be identified in AI models, mainly, due to the influence that the language, impregnated by gender dichotomies, present in the construction of algorithms and the learning of these machines. For Leavy, “the machine mainly learns by looking at the data with which it is presented [...] although a machine's ability to process large volumes of data can partially solve this problem, if the data is loaded with gender-stereotyped concepts, the resulting application of technology will perpetuate these biases” (p. 14).
At least since the 1970s, feminists have investigated the role of language in gender dynamics, as Leavy points out, “gender ideologies are still embedded in textual sources and result in machine learning algorithms that present stereotyped concepts of gender”. The use of the terms "men" and "women", for example, are associated with different functions, representations and naming: algorithms demonstrate the recurrent use of the term "family man", with no equivalent for a woman, at the same time, in that terms such as “single mother”, “working mother” and “career woman” serve to describe preconceptions about the social role of women; still, the masculine designation being used as universal, teaches the algorithms to perceive certain functions as assigned to men. Also, the descriptive terms associated with men and women vary greatly, the term 'girl' is used more than half of the time to refer to women, while less than 30% of the use of 'boy' is attributed to men; the term 'wife' is much more used than the term 'husband'.
In another example, in 2013, UN Women in association with advertising agencies released the search algorithms related to women in the Google search engine, in order to show what is displayed when women is searched on Google. When you type “women cannot”, the autocomplete features suggest “cannot drive”, “cannot be trusted”, “cannot be pastors”; or even, when searched for “women should” the surveys auto-complete with “should stay at home”, “should be in the kitchen”, “should be submissive”, etc. In 2016, Google confirmed the removal of suggestions related to the terms of its system, when typing woman or women, today there are no suggestions. However, the challenge remains: insofar as the algorithm that generates such contents continues to reproduce the most diverse prejudices. The 'autocomplete' feature tries to 'predict' users' thoughts from the data generated by the algorithms. As a result, auto-completion could, in fact, influence a search that the user did not even initially intend to perform. Several authors point to how these systems perpetuate gender stereotypes.
In this sense, in “Constructions of gender in the history of Artificial Intelligence” (1996), Alison Adam also presents a discussion of how ideas can be gendered and produced in the field of AI to be associated with notions of masculine and feminine. Adam argues that the reasoning and intelligence models of the AI field essentially involve two-order epistemological problems: who is the ideal knower; and, in terms of what can be known (ADAM, 1996, p.48). The general presupposition of epistemology that "S knows that P" as the only form of adequate knowledge assumes that all knowledge needs to be propositional, not being, it cannot be knowledge, which excludes practical skills and know-how from the validation as knowledge. For Adam, the symbolic field of knowledge of Artificial Intelligence remains in search of “the disembodied ideal of the 'Man of Reason'” (ADAM, 1996, p. 49). Problem solving systems, already in the first AI models, such as the “Logic Theorist” by Newell, Simon and Shaw, start from the presupposition of the ideal of male reason, insofar as:
“In itself we should not take for granted the idea that solutions to problems are things to be searched for. The idea of search is a very fundamental part of symbolic AI. Search techniques are based on the ideal Cartesian method of deduction, and this disguises the need to look at how other forms of problem-solving based on intuition (seen to be a less prestigious form of reasoning) or creative leaps could be represented where a search is not ostensibly part of the process” (ADAM, 1996, p. 49).
Adam concludes that the field of Robotics and Artificial Intelligence can face issues of situated and embodied knowledge, in which "intelligence is not seen as the individual's construction of a mental representation of the world, but rather an emergent phenomenon resulting from the individual's interactions with its environment” (ADAM, 1996, p. 51). The “COG” project developed by MIT, according to Adam, advances the hypothesis of situated knowledge for the field of Artificial Intelligence, as the system “was based on the hypothesis that intelligence at the human level requires gaining experience in the interaction with humans, as human babies do”.
Can women be cyborgs?
Recent researches that analyze gender roles and Artificial Intelligence systems indicate that the content considered as 'knowledge' in these systems maintains paradigms criticized for a long time by different feminists epistemologies: that the supposed neutrality, objectivity and universality of what we can categorize as knowledge hides criterias that favors a male view of the world, regarding what can be known and who can know. Furthermore, the kinds of problems and solutions these AI systems deal with still largely respond to the demands of a male-dominated world.
The barrier to the integration of women in the world of science and technology is not only related to their lack of access and participation, although it contributes to the diagnoses discussed here, but also lies in the alienation from what these systems seek to analyze, optimize, maximize and to know. Tools are still needed so that we can integrate a kind of 'feminine' knowledge, or even multiple experiences of the world, in what we consider as knowledge, so that, finally, women can integrate themselves into the imaginated new circuits and systems.
The image of the cyborg, as this hybrid cybernetic intelligent organism, of machine and living organisms, which is a real social creature and at the same time fictional, which produces an ambiguous notion of the natural and the constructed, is said by Haraway as the creature of a world 'post-gender', although Haraway dismisses the term ‘post-gender’ itself (HARAWAY, 1991, p. 150), insofar as it transgresses the dichotomous divisions that structure the conceptions of gender. Haraway claims that:
"A cyborg world can be about social lived and bodily experiences in which people are not afraid of their joint kinship with animals and machines, not afraid of their permanently partial identities and contradictory points of view" (HARAWAY, 1991, p. 154).
The provocation of the question 'Can women be cyborgs?' seeks to reveal, precisely, the masculine character and the model of Lloyd's 'man of reason' in the scope of Robotics and Artificial Intelligence.
The integration of the lived body with technology, in the sense that Paul Preciado describes as “productions of biotechnopolitical subjectivities” (2018), certainly makes room for the imaginative creation of forms of existence that transgress Enlightenment dichotomies and gender roles: the use of the contraceptive and day-after pill, takes away from women the deterministic reproductive role, the application of testosterone gives females a way of accessing the legitimized and valued attributes considered exclusively of male, the coupling of Dildo serves as a technology for co-opting power in a phallocentric society.
However, even so, it seems that our transformation into a cyborg reality, as described in the Haraway manifesto, still lacks a shift of the epistemological paradigm, in which our knowledge and imagination of the world is able to describe more multiple and diverse experiences, which have problems, demands and solutions different from those pre-established by the male reason.
Ferrando, in a questionnaire applied to students from the Department of Cybernetics at the University of Reading (England), in order to determine perceptions and representations of gender in the field of Robotics and AI, points out that the majority of participants in the research, which exceeds one thousand, associated the cyborg as male or neutral, but none related the image of the cyborg with the female (FERRANDO, 2014, p. 6). For Ferrando, technology “is not only performed, but first imagined” (FERRANDO, 2014, p. 7), it is the imagination that is culturally situated that informs what we intend to know, thus “If the genealogy of knowledge silently informing AI is reduced to a male legacy, social exclusivism and biological essentialism may be re-inscribed in its ontology, with the consequent risk that the difference characterizing robots may be assimilated in human-centric practices of assimilation; parallelly, it may turn into a stigma for new forms of discriminations" (FERRANDO, 2014, p 7).
“For this reason, employing critical frames such as Feminist Epistemology, the Philosophy of Sexual Difference, Critical Race Theory, Postcolonial Studies, Queer Theory, Disability Studies and Intersectionality, among others, is seen as crucial in the development of posthuman epistemologies informing the technological fields. Adopting such standpoints will allow humans to generate an emphatic approach, preventing them from turning the robot into their new symbolic other, and from falling into the dualistic paradigm which has historically characterized Western hegemonic accounts, articulated in opposites such as: male/female, white/black, human/machine, self/other” (FERRANDO, 2014, p. 16).
This leads us to realize that parity of participation and diversity in all fields of knowledge is not only an ethical commitment to egalitarian principles, but is also the way for the content of our knowledge and the problems we intend to solve with new technologies, encompasses a world closer to the real: that which is situated, partial, historical and of points of view. This feminist methodological commitment, it needs to be clear, does not imply the complete abdication of notions of objectivity, reality and specific criteria for scientific research; but it opens avenues for a constant debate of new paradigms, new worlds, and, above all, new ways of knowing the world that situate ourselfs.
References
ADAM, Alison. (1996). Constructions of Gender in the History of Artificial Intelligence. IEEE Annals Of the History of Computing, 18(3).
FERRANDO, Francesca. (2014)Is the post-human a post-woman? Cyborgs, robots, artificial intelligence and the futures of gender: a case study. European Journal of Futures Research volume, 2(43). doi: 10.1007/s40309-014-0043-8.
HARAWAY, Donna. (1995) Saberes Localizados: a questão da ciência para o feminismo e o privilégio da perspectiva parcial. Cadernos Pagu, (5), 07-41.
HARAWAY, Donna. (1991). Simians, Cyborgs, and women: the invention of nature. New York: Routledge.
HARDING, Sandra G. (1986). Whose science? whose knowledge?: thinking from women's lives.New York: Cornell University Press,.
JAGGAR, Alison M. (1989) Love and knowledge: Emotion in feminist epistemology,
Inquiry: An Interdisciplinary Journal of Philosophy, 32:2, 151-176.
LAZREG, Marnia. (1994). Women’s experience and the feminist epistemology. In: LENNON, Kathleen; WHITFORD, Margaret. (ed.) Knowing the difference: feminist perspectives in epistemology. New York: Routledge.
LEAVY, Susan. (2018) Gender Bias in Artificial Intelligence: The Need for Diversity and Gender Theory in Machine Learning. ACM/IEEE 1st International Workshop on Gender Equality in Software Engineering. doi: 10.1145/3195570.3195580.
LLOYD, Genevieve. (1985). The man of reason. Minneapolis: University of Minnesota Press.