A feminist approach to assessing AI, privacy and data protection in South Africa

December 5th, 2020

The replication of existing inequalities, development of new social injustices and unequal power dynamics impact the difference in experiences of AI and machine learning systems for marginalised groups. A feminist approach is used to assess the issues at hand beyond compliance for economic engagement, but rather the use of the technology within the context of social inequalities. Context is central to the understanding of what can be done to address the issues at hand from a gender perspective. 

Table 1: Conceptual framework: Feminist approach

A feminist conceptual lens
Data Feminism
Catherine D’Ignazio and Lauren F Klein.
7 guiding principles of data feminism
Examine power
Challenge power
Elevate emotion and embodiment
Rethink binaries
Embrace pluralism
Consider context
Make labor visible
Feminist principles of the Internet The principle of privacy and data that supports the right to privacy and full control over personal data and information online at all levels
Intersectionality
Kimberle Crenshaw, Patricia Hill Collins and Sylvia Tamale
An understanding of social interactions with technology that consider multiple inequalities and locating technology in the context of systematic oppression including racism, sexism, colonialism, classism, and patriarchy
Data Justice                Linnet Taylor; Data justice Lab and Global Data justice A way to center marginalised groups in datafied societies to recognise opportunities and respond to harms emerging from use of data in society that may harm groups in society.
Feminist internet ethical research practices A feminist guiding tool with consent, accountability, respect, privacy, safety and reciprocity  the main ethical concerns of this study

The feminist conceptual framework is build from various schools of thought to understand the issue and context at hand. Data feminism (see table 1)provides the following principles as guidance: examine power – the way it operates in the world; challenge power – to push back against these power dynamics and work towards justice; rethink binaries – challenge the gender binary and binaries that lead to oppression; embrace pluralism – bringing together multiple perspectives while prioritising lived experiences of the communities affected and focusing on local and indigenous knowledge;  consider context – locate this conversation in context to understand the unequal social relations; make labour visible and elevate emotion; and embodiment of value in multiple forms of knowledge. 

These guiding principles allow for critical engagement and centering of society concerning technology and the current laws. In centering society the focus is on its differences, challenging neutral approaches to law and technology. 

The feminist principles of the internet on privacy and data protection also form the underlying conceptualisation of this work. The right to privacy and full control of over personal data and information online at all levels is championed in supported in the principle. Practices in the public and private sphere where data is used for profit and to manipulate behaviour online is rejected. In the principle, surveillance from private and state actors is paid attention to given its historical use as a tool of control and restrictions of women’s bodies, speech and activism.

The experiences of inequality in society are different. Intersectionality allows for us to look at the layers of inequalities based on the different spaces we occupy. Kimberle Crenshaw highlights that intersectionality allows us to see inequality of gender experienced at various points including race, location, class  and sexuality. Furthermore, Patricia Hill-Collins shows that there are domains of power that we exist in at different times that shape our experiences of opportunities and inequalities at varying intersectionalities. The four domains of power Hill-Collins identifies include the structural domain – the design and focus of the law; disciplinary domain – how things are done; hegemonic domain – norms that drive the space; and the interpersonal domain – how we relate to each other. As  Sylvia Tamale writes, ‘while Africans are adversely affected by enduring legacies of colonialism and its convergence with racism, our positioning within diverse social categories based on gender, ethnicity, class, sexuality, disability, religion, age, marital status etc. means we experience oppression differently.’  The intersectional approach allows for an understanding of gender responsive laws that consider multiple inequalities and locating technology in the context of systematic oppressions including racism, sexism, colonialism, classism, and patriarchy.  

Feminist research is interested in ways in which the work it does contributes to how technology can be used for transformational change in society for women, gender diverse and vulnerable groups based on class, sexuality or ethnicity. The research is interested in ensuring data justice. Nancy Fraser’s work on abnormal justice challenges us to rethink justice by focusing on ‘what of justice, who of justice and how of justice as a disruptive way of thinking of justice.’ 

A data justice approach acknowledges the complexity of the new technology systems and how they can be used to discriminate, discipline and control; take into account the positive and negative potential of these new technologies; and; make use of principles useful across varying contexts. The data justice approach privileges social conditions and lived experiences of those who are subject to domination and oppression in society. Our entry point is not the data system itself but rather ‘the dynamics upon which data processes are contingent in terms of their development, implementation, use and impact.’ Our starting point is the lived experiences of data systems, the perceptions of injustices and the subsequent point of how to move beyond this. 

Methodology: Interviews and targeted survey

Guided by feminist epistemologies, data has been sourced through secondary and primary data which allow for multiple perspectives of knowledge. The secondary data draws from literature and an assessment of current legislation related to digital rights – specifically privacy and data protection. Through a process of mapping stakeholders from the data and a snowball methodology, a qualitative methodology of interviews was implemented with ten individuals from the technical, academic, and legal community (see list of participants in acknowledgements). 

A targeted survey was used to engage activists working in the gender and sexual justice community. In line with the data justice approach, the focus of gender and sexual justice activists is based on gaining insights on gendered harms from those whose work is focused on gendered inequality in society therefore centering marginalised communities. . The survey was a tool to gauge awareness and concerns of the right to privacy and data protection considering AI uptake in South Africa and is in no way represantative. In total, 25 participants engaged with the survey. The participants came from diverse work spaces such as research, media, human rights, and sexual reproductive health rights. These individuals work across women and gender diverse communities which allowed for an intersectional approach to understanding the issues at hand and multiple forms of knowledge. 

Ethical considerations for this study were based on feminist internet ethical research practices. In thinking of consent in both interviews and surveys, the purpose of the study clearly explained the goals and purposes of the information provided for the study. In thinking of accountability in ethical practices, the responsibility of the study lay with the researcher in communicating with the participants any harms that may emerge. Participants retained the right to withdraw consent within a given time period.

In conclusion

A feminist approach allows one to ask questions of who is being represented and by whom; whose interests are being centered; why this discussion is important and how it is taking place, which allows for criticism of power and how data itself can be used to ensure justice in society.

Acknowledgment to research participants – the following research participants consented to being named as having been part of this research process. 

Interviewees Survey participants
Pelonomi Moiloa Technical Community
Kerryn Gammie – Technical Community
Daniel Mpala – Technical Community
Gabriella Razanno – Policy
Dr Rachel Adam – Policy
Tina Power – Legal
Mutondi Mulaudzi – Legal
Amanda Manyame – Legal
Alice Piterova – Legal/Technical
Tobias Schoonwater – Legal/Technical
Tshegofatso Senne
Nyx Mcclean
C Alves Nascimento
Nonhlanla Chanza – Governance Expert and Parliamentary Officer
Subisiso Khuzwayo
Otto Saki – Lawyer, Data Privacy
Tiffany Mugo
Gaopelelwe Phaeletsile – the founder of Abortion Support South Africa
Tigist Sheweraga Hussen – project coordinator, Feminist Internet Research Network, APC