Gender Matters: “Socializing” Artificial Intelligence

Abstract

“Why are virtual assistants always female by default?” Given the feminized empathetic machines such as Alexa, Siri, Erica, Sophie, or Google Home, substantiates the conventional belief that women are more suitable for the role of assistants or companions. One may have to understand that these machines are none other than the product of codified opinions that gendered the roles of virtual assistants. This context needs to be perceived through the radical feminist theories to elucidate and challenge the gendered epistemology in the field of Artificial Intelligence. These theories help to analyze how male-based authority, power, and politics play a crucial role in structuring and socializing the bodies into gendered roles and behaviors. Secondly, Feminist scholars of science and technology find such Gender Problems in Artificial Intelligence as an ethical issue. In this context, the paper affirms that this tech age diffuses the stereotypes, produces, and reproduces patriarchy, which needs to be identified and sorted out in order to provide a gender-inclusive space to construct these virtual robots.

Introduction:

Empathetic machines are the one that can detect and respond to human emotions (Grant, 2017). Chatbot, like Sophie, Erica, or Alexa, expresses a feminine character that pleases the consumers. On the other hand, the rescue robots or the one which can Parkour, from Boston Dynamics, are masculine in shape and form. The patriarchy is corrupting even the robot. This context should be viewed through the lens of the theory of socialization. Sociologists define it as a process of inheriting and teaching norms, customs, and ideologies, providing the human being swith the skills and habits necessary for participating within the society. In other words, it is the means through which social and cultural continuity are attained. 

Peer groups, parenting, education can media are a few examples for the socializing process, which shapes one's personality. For instance, cultural socialization refers to the parenting technique that teaches one's children about their racial history or heritage and, sometimes, it may lead to developing pride. Likewise, as per the context of this paper, gender socialization is the medium through which one can learn behavior and attitudes considered appropriate for the given sex (Wharton, 2005). In the sphere of Artificial Intelligence, this provides scope for innovation, radical approach, deviant behavior, or cultural shock, however ender up in producing the stereotyping robots that reflect the sexist and patriarchal nature of the society. What is the factor behind this theory and practice of this gender divide in Artificial Intelligence?

Theoretical Approach via Feminist lens

Why AI has gendered underpinnings? AI influences how people are perceived and treated in society. However, it has a reciprocal relationship, i.e., the technology adapts and reproduces gender relations, which are then repetitively reproduced. As Wajcman (2010) points out, “Gender relations can be thought of as materialized in technology, and masculinity and feminity, in turn, acquire their meaning and character through their enrollment and embedded in working machines.” Similarly, Judit Butler theorizes how gender is embedded in temporal repetitions, which re-enact and re-experiencing the meanings that are already established in the society. Moreover, the repetitiveness is amplified by the growing magnitude of AI development across the globe.

The famous feminist discourse on Science and Technology has been looking at the reciprocal relationship between AI and gender relations upholds the argument of Alan Turing (1950), which emphasizes that computers function according to the principle of imitation but also learn new things. Halberstam infers the parallel argument that gender is a “learned, imitative behavior that can be processed so well that it comes to look natural.” (Haberstam, 1991). Lauren Wilcox throws light on the intersectional approach and recognizes the impact of gender binary and colonialism on AI, which attempts to create, fix, control, and sustain the hierarchy. The process was termed as “infrastructural imperialism” (Vaidhyanathan, 2011). The argument also encompassed the racial approach, which stresses the socio-political relations that reproduces power structure that controls bodies. (Wilcox, 2017), Keyes discusses the consequences of such binary relations and argues that it inevitably discriminate against trans people and others. (Keyes, 2018)

Thereby the AI is entrenched in gender relations and explicitly visible in the hominoid robots. They uphold gender stereotypes through the power of language and naming. They construct the concept of female robots as a faithful aid of humankind, which exists only to assist. They have been named as “hey Siri,” “Hey Alexa,” to create what Judit Butler defines as power dynamics (Adam, 1995). The feminine voice of the AI is associated with servitude and power disparity, which has damaging implications. It enables surveillance, and further domesticate the feminine persona through promoting “digital domesticity” (Woods, 2008).

Practical Reasons

While going through interviews from techies, they tell the practical reasons behind this Scenario. Ivana Bartoletti, the founder of a leading AI network, stated that “An algorithm is an opinion expressed in code. If it is mostly men develops, then the results will be biased. You are teaching the machine how to make a decision”. There were assumptions and expectations that AI will reduce the bias, especially during the recruitment, and the present proved the reverse. The US tested robots for doing facial-analysis programs, and to their surprise, the machines picked men over women. Likewise, the Google algorithm was found to associate the word “nurse” with women automatically. An AI program driver at PwC says that the majority of the machine learning engineers are men, computer programming was historically the realm of war. Initially, more women have been recruited in AI development teams. Likewise, women occupied the maximum space in the codebreaking task. However, in the following decades, the programming shifted from the “low-status feminized task,” to a job that as seen as the power center of the corporation and state, which eventually edged out the women. In the aftermath, 66.7 virtual assistants are female. According to the study conducted by Tyler Schmoebelen, even when the creators refer the robots as “it”, people can still assume it with a gender (Schnomebelen, 2016).

Why does Representation matter?

Representations may seem to reflect society merely. Nevertheless, it is the place of many perspectives, and it is the representation that maintains and creates it. This AI scenario does not stop with the virtual world but has an impact on a day-to-day basis. For example, it increased the surveillance of minority communities or making it harder to get loans. The homogenized nature of the decision-making will easily label suspicious people and gives a default answer on which group of people knows it very well. The group may blackout a certain race, language, and gender. Gives an example, wherein a specific job dominated by a gendered stereotype, dominance causes a disadvantage to a specific group of people, which is ignored by society. Secondly, there are intense ethical issues and reliability. Sociologists argue that gendering robots are based on the “common sense” of the technologists, which might cause users to demonstrate abusive behavior towards the robots. (Nomura, 2017)

Discussion

As discussed earlier, AI is an effective socialization tool that imparts about the role of women and people who are gendered to respond to the demand. While the commanding AI has a masculine voice, the virtual assistant is feminine to please the consumers. This issue has been perceived as both ‘gender problem’ and an ‘ethical issue.’ Rather than making a call that robot does not need to be gendered, one could create a conducive sphere that fosters a culture of diversity and inclusiveness at the workplace and make sure that diverse points of view are inferred in the designer and monitor these solutions. Having business analysts, ethicists, and experienced designers who could bring diverse skills and gender diversity could create a customized robot that serves the interests of all without bias. As the application of AI continues to expand into multiple fields, it is essential to take note of the implications of gendering the AI and use the technology wisely to reduce the harmful gender bias.

Bibliography

1. Adam, A. (1995). Artificial Intelligence and Women's Knowledge. Women's Studies international forum.
2. Grant, B. (2017). Empathy machines. Meduia International Australia.
3. Haberstam, j. (1991). Automating Gender:Postmodern Feminism in the Age of intelligent Machine. Feminist Studies, 439-460.
4. Keyes, O. (2018). The Misgendering Machines. CSCW Arcives. New york.
5. Nomura, T. (2017). Robots and Gender. Gender and the Genome, 1(1).
6. Schnomebelen, T. (2016). The Gender of artificial intelligence. The Medium App.
7. Vaidhyanathan, S. (2011). The Googlization of Everything: Why we should worry. University of Califonia Press.
8. Wajcman, J. (2010). Femnist Theories of technology. Cambridge Journal of Economics, 143-152.
9. Wilcox, L. (2017). Embodying algorithimic war: gender race and the posthuman in drone warefare. Security Dialigue, 11-28.
10. Woods, H. S. (2008). Asking more of Siri and Alexa: Feminine Persona in servie of survillance capitalism. Critical Studies in Media Communication.

Cite this page