Different Stories are Possible: On Data, Feminism and Inclusion

“Simians, Cyborgs and Women” -Donna Haraway

The scientific and mathematical biases that occur in Big data and Machine learning processes are well-known and widely studied. Big data’s idealistic quest to uncover a truth, as conveyed by an extensive, if not complete, dataset, is often challenged by statistical bias, anomaly and bad algorithm design. The more big data is used to inform decision-making in various industries and social contexts, the more that it becomes apparent that the critical questions in big data are beyond mathematical or scientific. One of the questions that the puzzle of big data is concerned with building inclusive, feminist and transparent machine learning processes.

When speaking of transformation, especially a transformation that is data-led, it is important to ask the question “For whom?”. For whom is the transformation intended? Who gets to be involved in the transformation process? For whom is the transformation oppressive, counterproductive or harmful? The word ‘inclusive’ is often tokenized within institutions. For example, ‘inclusive design’ becomes a code word while the consideration of inclusion may be minimal. Sara Ahmed’s work continuously touches upon this misleading nature of ‘appearing to be inclusive’, pointing that “Equality and diversity can be used as masks to create the appearance of being transformed”.

But shouldn’t science remain objective? And shouldn’t data remain a clear-cut decisive evidence, operating away from the sensitivities of social nuances?

The answer to this lies within Harawayian thoughts where she explains, as she examines the Natural movement in the 30s, that “Machines are maps of power, arrested moments of social relations that in turn threaten to govern the living”. Haraway’s critique of technological determinism stays relevant to date, as can be seen with Facebook and their insistence on the ‘objectivity’ of their algorithms and that bias, racism and exclusionary filter bubbles are the sole result of social processes. David Banks traces Facebook’s technological determinism in depth in this blog post .

So, in the context of big data, how can we create true, non-performative and inclusive data-driven transformations? The increasing participation and reliance of big data in decision making across disciplines makes this question not only inevitable, but urgent.

This year at Mozfest 2018, a horizon of feminist stories were illuminated through big data.

Feminism-oriented data comes into question for replacing the older paradigm of data-driven feminism. Mozfest, under the hood of loving the internet, has organized many spaces of dialogue: queering the internet, digital inclusion, data decentralization, openness, literacy and security. Feminism-oriented data comes at the intersection of all these spaces.

In this blog post, I aim to highlight some of the work that goes beyond the feigned representation of diversity.

1. Feminist Data Sets ~ Caroline Sinders

Link: https://carolinesinders.com/work/#/feminist-data-set/

Caroline Sinders aims at diversifying the input of Machine-Learning systems. Stemming from the well founded “garbage in — garbage out” principle, skewed datasets will reproduce skewed perceptions of society. According to the project’s description, feminist data “can be artworks, essays, interviews, and books that are from, about or explore feminism and a feminist perspective”. Here, a diverse input is taken into account in the data collection process. Caroline’s project lies on principles of co-creation: in principle, it’s a call to action, inviting people who work with data to submit their feminist data sets.

More theoretical insight into Data Feminism can be found in this free open source book Data Feminism

2. Data Drag Project

Link: https://insta-stalker.com/post/BqNMZ0ChJZZ/

Theoretical work on representative and nonconforming views on data is also lacking. In their project Data Drag, Stephanie Ouillon and John Phillip Sage elaborate on the connection between drag, gender performativity and data movements such as the quantified self.

3. Zoyander Street

Link: https://zoyastreet.wordpress.com/

Zoyander Street uses data to tell stories and allow users to have conversations about inclusion, diversity and nonconformity they usually would not have. In essence, Zoyander uses fiction to influence realities.

Zoyander’s work was also exhibited at the MozillaFest Open Art+Data Exhibition.

These examples are only tools of illustrations; they are neither exhaustive nor complete. They mainly illustrate that data and machine learning can be used to enable new modes of inclusive and feminist storytelling. In the context of data-driven transformation, it is important to acknowledge that transformations are not only about digitizing the social with a marginal consideration of issues such as diversity. Rather, diversity must be centered and embedded in the transformation work. Collecting feminist data, creating feminist narratives, cultivating feminist data theory and involving actors of feminism are all tactics that can be employed to put diversity at the center of digital transformations.

The Digital Society School is a growing community of learners, creators and designers who create meaningful impact on society and its global digital transformation. Check us out at digitalsocietyschool.org.