Why data justice can be the starting point in development

Why data justice can be the starting point in development

It has become worthless to point out that algorithmic systems increasingly permeate the social sphere and day-to-day life. ‘Smart’ technologies, machine learning and Artificial Intelligence are now an integral part of how societies are organised. The way data is generated, collected and used, then, has become an increasingly prominent issue.

How do we come to understand the world? What services we are able to access? Where we can go? In which way we are governed? It’s likely all feature data practices that shape the terms and conditions for our participation in society (Dencik, Hintz, Redden & Treré, p. 873).

Datafication of devolopment

The widespread adoption and use of ICTs in the context of development is seen to bring opportunities for improving the people’s lives (Qureshi, p. 202). Complex socioeconomic problems such as poverty, environmental safety, food production, security and the spread of disease are automated and packaged as mathematical and engineering problems.

But, as Harcourt (2015) describes it, for data systems to be scalable to generate sufficient meaning, there is a necessary trend to reduce social identities, mobilities and practices to data. 

The ‘datafication of development’ can be described in terms of a growing volume, velocity, variety and visibility of data, with greater use of new forms and streams of data in decision making (Heeks, 2018 cited by Heeks, Shekhar, p 993).

A tendency to turn vast amounts of activity and human behaviour into data points, that risk to be managed and sorted as abstractions without a clear understanding of the embodied power relations and social effects produced by those activities (Monahan, 2008; Costanza-Chock, 2018 cited by Metcalfe, Dencik).

Data-driven decision-making

With these developments, we are confronted with a significant shift in governance. Policymakers and researchers worldwide are taking advantage of the increasing availability of digital data – what the UN has termed the ‘data revolution’ (DataJustice Project).  

From being data-informed to being data-driven, the turn to data collection and algorithmic decision-making is not simply a question of quantity. A quality shift is required. These data sources have ethical, political and practical implications for the way people are seen and treated by the state and by the private sector (Kitchin, 2016 cited by Taylor, p. 1).

Although this is often welcomed as ‘revolutionary’ in its potential for enhanced efficiency, security and innovation, it sure does increase concerns with the societal implications (DataJustice Project).

The dark side of the ‘revolution’

“This brave new world of solving pressing problems through machine learning has several dark sides” (Qureshi, p. 201).

It is undeniable that data has been essential to both delivery and measurement of the Sustainable Development Goals. Unlike the predecessor Millennium Development Goals, the SDGs appear less problematic when it comes to implementing policies and procedures directed at achieving these goals, thanks to these new data (Qureshi, p. 204, 205). As mentioned, ICT use seemed to present great opportunities for advancing people’s rights and representation.

But what about those who do not have access to or use ICTs? Their data cannot be extracted nor analyzed, they remain under-represented in digital datasets and become invisible to policymakers and agencies working to implement development efforts. On the other hand, refugees’ every move is tracked and their behaviours analysed, producing modes of mapping that are ‘colonial’ – difficult to criticize since the assumption is the data analysis’ ability to predict and optimize migration and its implications (Qureshi, p. 205)

The transition to greater visibility creates informational and power asymmetries in the form of surveillance and the potential for manipulation and exploitation (Dencik, Hintz, & Cable 2016). Heeks and Renken (2016) refer to the disbenefits developmental impacts of datafication which include 

  • growing surveillance and loss of privacy, 
  • the capture of development gains by private corporations, and 
  • growing inequalities

Yet somehow the ‘data revolution’ is always presented as completely beneficial for the Global South. However, a growing body of research on algorithmic injustices has shown how ML automates and perpetuates historical, often unjust and discriminatory, patterns (Birhane, p. 2). Datafication not only introduces and digs into key questions about how an individual’s data is collected, stored, or algorithmically processed. The issue is rather how data-driven decision-making is part of an economic and political agenda that seeks to stigmatize, marginalise, and exclude certain communities of people. (Metcalfe, Dencik). 

Data-driven discrimination, as Taylor (2017) refers to, is “advancing at a similar pace to data processing technologies, awareness and mechanisms for combating it are not” (Taylor, p. 1). Unfortunately, there is no awareness of how new data technologies may not be neutral in terms of access, use or impact (Taylor, p. 1).

All these transformations happening across governments and civil society may require a wider framework for understanding what is at stake (DataJustice Project). Many authors call for data justice and increase protection of citizens’ rights (Dencik et al., 2016; Taylor, 2017; Taylor, 2019; Heeks & Renken, 2018; Masiero & Das, 2019 cited by Qureshi, p. 202).

Social justice in an age of datafication 

The power of data to sort, categorize and intervene has not been explicitly connected to a social justice agenda by the agencies and authorities involved (Taylor, p. 1). Many solutions offered 1) revolve around technical solutions and 2) do not center impacted communities. Such work can be part of the remedy, but it cannot be enough. Automated and standardized solutions to social issues often fail to understand their complexity and provide a “false sense of safety” (Birhane, p. 2). 

An increasing emphasis is being placed on the fact that data processes are not ‘flat’. Inequality in the distribution of benefits and harms is already uneven between countries: asymmetries in information flows, social structures and political power but also embedding of racial, gender and other biases into AI systems (Data Justice Project website). 

Regulation of data needs to engage more explicitly with questions of “power, politics, inclusion and interests, as well as established notions of ethics, autonomy, trust, accountability, governance and citizenship” (Dencik, Hintz, Redden & Treré, p. 874).

A call for Data Justice

Data and data-driven technologies are recognized as not neutral artifacts anymore, and what is at stake now cannot be simply captured “by simple binaries such as efficiency vs. privacy, or good vs. bad data” (Metcalfe, Dencik). Instead, data is seen as something that is necessarily understood in relation to other social practices (Dencik, Hintz, Redden & Treré, p. 873).

The new framework, that makes an explicit link between data e social justice – ‘data justice’ – refers to the “fairness in the way people are made visible, represented and treated as a result of their production of digital data” (Taylor, p. 1). The ideal notion of justice is not enough anymore to question the implications of unaccountable decision-making created by datafication. Establishing a notion of data justice is vital for ethically desirable developments in datafied and datafying societies. 

The approach to data justice takes the technology and the data processes as the entry-point for highlighting social justice issues. First, you identify and expose how data systems are affecting marginalized communities, their cultural and political participation, and access to fundamental rights. Secondly, rather than merely technological, engaging with data justice involves the active collaboration between different groups and movements within the civil society in articulating both problems and solutions (DataJustice Project). 

Such an approach will hopefully (re)politicize data and demonstrate its relevance to social justice issues (DataJustice Project), broadening the terms of the debate. In other words,

“data justice is a lens through which we can understand the relationship between data and social justice, to critique the political agenda that governs datafication and allows us to understand how data contributes to structural conditions that continue or create new injustices” (DataJustice Project).

Data is seen as a way to challenge the dominant understanding of the world, and create possible counter-imaginaries (Dencik, Hintz, Redden & Treré, p. 875).

What Data Justice means 

The notion of data justice connects different approaches, interpretations and concerns of the interaction between data and social justice. Our attention has moved towards the nature of transformations in governance and structural inequality, highlighting the disparity of experiences of data across different groups and communities. 

Data justice can be conceptualized in (at least) three main approaches (Taylor, p. 6):

  1. The first one relates to how data is used for governance and how these practices make asymmetries in the representation and power of data explicit (Johnson, 2018). Johnson connects data justice primarily to open data, asking how database design can better incorporate anti-discrimination principles
  2. An alternative interpretation calls for data technologies to provide greater distributive justice. Heeks and Renken (2016) focus on the question of how data should be distributed to achieve fairer access, participation and representation
  3. Thirdly, an approach that calls for the introduction of ‘data justice’ terminology to describe resistance to government surveillance based on principles of social justice (Dencik et al., 2016). Their framing focuses on social activism, and how to protect the work of activists working towards social justice. Under which conditions data should not be distributed?

New issues keep arising that research needs to address. This shows the complexity of bringing the notions of data and social justice into dialogue. It would make sense to position the notion of ‘data justice’ as an end-goal for foregrounding both the politics in data as well as the politics of data against inequality, oppression and domination (Dencik, Hintz, Redden & Treré, p. 876). 

3 pillars of Data Justice 

A ‘data justice’ overarching framework is possible. As articulated by Taylor (2017), it has three main components: visibility, engagement, and nondiscrimination, questioning how it is possible to balance and integrate the need to be seen and rightly represented without losing autonomy. 

Visibility deals both with informational privacy and representation, taking into account the “need to be represented but also the possibility to opt-out of data collection or processing” (Taylor, p. 8). Engagement with technology is the second pillar, critically addressing the freedom not to use particular technologies. It sustains the power to determine one’s visibility by controlling the terms of one’s engagement with data markets. The third pillar is nondiscrimination, involving the capacity to recognize and oppose big data biases (Taylor, p. 9). 

These possibilities may seem to require a certain degree of user expertise. But how mentioned in a previous blog post, data literacy is important if it aims at creating an informed dialogue between people and policy on data governance. Asking people to be responsible for their data is a way to divert attention from the lack of regulation on many shady data collection and use.

The three concepts represent a way to think about data that goes beyond particular applications, and instead to address technologies as they relate to human needs (Taylor, p. 9). Taken together, they formulate a foundation for protecting individuals from data trafficking, misrepresentation and exploitation (Taylor, p 5), without being obstacles to innovation. Evolution is technology is unstoppable and desirable, but we should be able to determine our interactions with it (Taylor, p 12). 

This framework exposes the transformations and implications of datafication that include ethics, law, sovereignty, and decoloniality amongst others. It involves a wider range of stakeholders in asserting the nature of both challenges and possibilities that may not be about data (Dencik, Hintz, Redden & Treré, p. 876) and it relies on the empowerment of impacted communities to effectively engage with decision-making (GPAI ToR)).

Conclusions 

Data justice helps frame the necessary questions and offers a roadmap toward further analysis, conceptualizing what we can and should about data governance. The principles set out here would need to be translated and rooted in local experiences, facing different challenges depending on the location of the discussion. 

By applying a data justice approach every legal and social system would need to attempt establishing coherent positions on what the boundaries and systems of governance should be concerning data for their own. And it should be doing this over and over because data justice has no answer to the problems. 

Establishing governance of data on the national level is necessary, but not enough. At the global level, we need to make those approaches align among them and with what people will value in the future.

Personal reflections 

I was amazed by the power of cooperation within the group. Being able to confront your ideas with someone else always gives space to create something completely new and unpredictable, and 98% better than the idea of one only person. I learned how to manage a blog and all its urgencies. Especially, I had the chance to explore SEO and its benefits for my writing. Being rules that aim at simplifying the reading experience, they can be applied any time you approach a topic, whether it is for work or academic purposes. Making something more readable does not mean losing the complexity; but rather finding a way to get straight to the point as much as possible. 

The use of a proper tone of voice – active, engaging – while I was managing complex information was not easy to find. A good exercise for me, daily involved in communicating via social media e shorter posts where it is crucial to find the right words and strike the right note. 

I started this journey knowing that I wanted to better understand, and explain to readers, how data works and how it does exclude different groups of people – for example, people that for several reasons cannot be traced (because they have no documents or no home), but also other groups of people that are being excluded and discriminated against by digitalization. Across the weeks and the posts, I have been able to develop a fil rouge that has found a proper space for reflection with this last article, where I try to identify a bigger picture where to frame all the singular themes I touched.

Bibliography

Birhane A., “Algorithmic injustice: a relational ethics approach”, 2021, Patterns 2, February 12, 2021

DataJustice project, “A conceptual framework for approaching social justice in an age of datafication 

Dencik L., Hintz A., Redden R., Treré E., “Exploring Data Justice: Conceptions, Applications and Directions”, 2019, Information, Communication & Society, 22:7, 873-881

GPAI (Global Partnership on AI’s), “Advancing data justice research and practice – Terms of Reference 04/06/2021”, 2021  

Heeks R., Renken J., “Data justice for development: What would it mean?”, 2018, Information Development, Vol. 34(1) 90–102

Heeks R., Shekhar S., “Datafication, development and marginalised urban communities: an applied data justice framework”, 2019, Information, Communication & Society, 22:7, 992-1011

Metcalfe P., Dencik L., “The politics of big borders: Data (in)justice and the governance of refugees”, 2019, First Monday, Volume 24, Number 4 – 1 April 2019 

Qureshi S., “Why Data Matters for Development? Exploring Data Justice, Micro-Entrepreneurship, Mobile Money and Financial Inclusion”, 2020, Information Technology for Development, 26:2, 201-213

Taylor L., “What is data justice? The case for connecting digital rights and freedoms globally”, 2017, Big Data & Society July–December 2017: 1–14