Coloniality and frictions: Data-driven humanitarianism in North-Eastern Nigeria and South Sudan

Abstract

It is now over a decade since the proclamation of a humanitarian ‘data revolution’, with the rise of ‘innovation’ and the proliferation of ‘data solutions’ rendering data-based humanitarianism an important area of critical investigation. This article contributes to debates within the field by exploring the role of data in the provision of humanitarian assistance within camps for internally displaced persons (IDPs) across north-eastern Nigeria and South Sudan. It draws on qualitative interviews carried out with humanitarian practitioners specialising in data and information management, as well as with camp residents and stakeholders located in each region. The analysis focuses attention on the ways in which epistemic injustices have been further perpetuated by the ‘data revolution’ due to the intensification of paternalistic dynamics associated with the coloniality of humanitarianism. It shows how a logic of extractivism structures the humanitarian data ecosystem, while also generating a series of tensions and disagreements. Data-driven humanitarianism, the article concludes, is characterised by recurring colonial dynamics as well as intensified frictions that bring epistemic injustices into sharper focus.

Introduction

It is now over a decade since the proclamation of a humanitarian ‘data revolution’ (Meier, 2012; Pearn et al., 2022), with the rise of ‘innovation’ and the proliferation of ‘data solutions’ provoking a range of urgent calls for the assessment of changing data practices in the sector (e.g. Burns, 2015). This article contributes to debates within the field by exploring the role of data in the provision of humanitarian assistance in camps for internally displaced persons (IDPs) across north-eastern Nigeria and South Sudan. In both contexts, the collection and use of large-scale quantitative data are well established, including through regular needs assessments and baseline data produced by international agencies and NGOs. Biometric data is also widely used, with basic provisions distributed using e-ID cards in both contexts. Data-driven humanitarianism is therefore embedded in IDP provisions both in Nigeria and South Sudan, raising questions about the risks and injustices that the digital turn generates in situations structured by longer-standing hierarchies of ‘humanitarian reason’ (Fassin, 2012).

In addressing such questions, we draw on qualitative research carried out as part of a multi-disciplinary international project focusing on camp-based internal displacement in north-eastern Nigeria and South Sudan. ¹ Our research team included practitioners from an international agency involved in data collection and humanitarian operations, as well as academics from the UK and from each of the regions under examination (Squire et al., 2022). As a team, we were cognisant of the disparities between us as colleagues and sought with varying degrees of success to find ways to invert the ‘data pyramid’ between the global North and global South (see Abimbola et al., 2021; Arora, 2016). As authors, we were based within the UK during most of the project, which coincided with several COVID lockdowns. Our role within the team has been to integrate the various dimensions of the research, and to provide intellectual and analytical support to the project team. In this article, we draw on findings from the first phase of the project research in Nigeria (Fayehun and Àkànle, 2022) and South Sudan (Logo and Jones, 2022), which involved semi-structured interviews during March–April 2021 with a total of 21 camp residents, ² and with 16 stakeholders involved in the local provision of protection and assistance. ³ We also draw on interviews conducted between April and July 2021 with 42 humanitarian practitioners specialising in data and information management from a range of international organisations, including donors, UN agencies and NGOs (Trigwell and Hirani, 2022). ⁴

The overarching argument that we advance in this article is that data-driven humanitarianism perpetuates epistemic injustices due to the intensification of paternalistic dynamics associated with the coloniality of humanitarianism. The analysis shows how the logic of extractivism structures the humanitarian data ecosystem, while at the same time generating a series of tensions and disagreements that we conceptualise as frictions over coloniality. Our argument is developed in four sections following this introduction: Section 2 develops the concept of epistemic injustice, while the subsequent three sections explore coloniality and frictions within the humanitarian data ecosystem. Section 3 draws on interviews with humanitarian data experts, to identify the ways in which data-driven humanitarianism involves various forms of extraction that are based on a hierarchical, paternalist logic. Section 4 highlights key areas of friction that such practices provoke. It draws on interviews with camp residents and stakeholders located in camps across north-eastern Nigeria and South Sudan, to emphasise the ways in which paternalistic relations of coloniality are subject to contestation. Section 5 goes on to highlight the continued limitations of accountability embedded within data-driven humanitarianism and the way this further reflects paternalistic relations of coloniality. Data-driven humanitarianism, the article concludes, is characterised by recurring colonial dynamics as well as intensified frictions that bring epistemic injustices into sharper focus.

Epistemic injustice, coloniality and paternalism

This article focuses on the perpetuation of epistemic injustices within the humanitarian field, to explore how the ‘data revolution’ involves an intensification of paternalistic dynamics associated with the coloniality of humanitarianism. Scholars have explored the intersections of social justice and processes of technological development from a range of angles over recent years (e.g. Dencik et al., 2016; Dencik and Sanchez-Monedero, 2021; Heeks and Renken, 2016; Heeks and Shekhar, 2019; Newman, 2015; Ruberg and Ruelos, 2020; Taylor, 2017). Our emphasis on epistemic injustice arises from concerns regarding the injustices arising due to the oversight of subjugated IDP voices and the failure to recognise the knowledge and capacities of those with lived experiences of displacement in decision-making processes (cf. Squire et al., 2021). Such concerns are informed by the work of Gayatri Chakravorty Spivak (1988), who highlights the epistemic violence that occurs when those in positions of power speak on behalf of others. Miranda Fricker defines epistemic injustice both in terms of inequalities of access to epistemic goods such as information, as well as in terms of the devaluation or neglect of a person or group as having the ‘capacity as a knower/subject of knowledge’ (2010: 5). Translated for our purposes, epistemic injustice can therefore be understood in terms of the production of IDP data by and for humanitarians, rather than by and for displaced persons themselves. This involves the devaluation of IDP knowledge and the failure to address people in situations of displacement as subjects with the capacity to know and make decisions about the production and use of data.

Epistemic injustices have a profound effect on the lives of displaced persons, generating individual and social harms while enacting epistemological closures through which IDPs are excluded from processes of knowledge production (cf. Milan and Treré, 2019). Our interest in this article lies in the ways in which epistemic injustices are generated by the intensification of paternalistic dynamics of power through data-driven humanitarianism. Scholars within the field have highlighted the dynamics of care and control that are embedded within humanitarianism (Rothe et al., 2021), as well as the ambivalent dynamics of power that are enacted through digital aid (Weitzberg et al., 2021). Our focus on paternalism refers to processes of subordination, whereby the freedom and responsibility of subjugated subjects is restricted. This is based on a logic of care, which assumes a subject (or object) that lacks capacity or competence (‘the IDP’), as well as a subject that is confident in their ability to make a positive difference (‘the humanitarian’). As Michael Barnett highlights, the latter ‘often possesses not only compassion and confidence, but also a sense of duty, responsibility or obligation to act’ (2016: 136–137). It is precisely through these paternalistic dynamics, associated with processes of professionalisation, that humanitarian expertise generates epistemic injustices and ‘the silencing of local voices’ (Barnett, 2016: 141). We suggest that the digital turn in humanitarianism, based on the professionalisation of the sector and the growing influence of data science expertise, plays a key role in intensifying the logics of paternalism.

In this article, we consider the relation of paternalistic dynamics to what we call the recurring coloniality of humanitarianism. Our analysis draws here on scholarship surrounding the ‘coloniality of power’, which refers to racialised social categorisations and epistemic logics that endure beyond colonialism as a historical social formation (Quijano, 2007, 2008). Scholars of migration have shown how unequal mobilities are governed in terms that represent an enduring global ‘colour line’ (Pallister-Wilkins, 2022), as well as ongoing structures of racial capitalism (Gutiérrez Rodríguez, 2018). Moreover, they have shown how a logic of extraction sits at the core of data-driven forms of migration governance (Aradau and Tazzioli, 2020). Madianou (2019) describes this as a form of ‘technocolonialism’ that reinvigorates longer-standing relations of colonial dependency. We do not focus our analysis here on territorial colonialism and its various historical manifestations or forms of domination (see Mignolo and Walsh, 2018). Rather, we focus attention on ‘recurrences’ of coloniality that emerge in the contemporary context (Lemberg-Pedersen, 2019; see also Lemberg-Pedersen and Haioty, 2020). This enables us to highlight how data practices are shaped by colonial legacies (Couldry and Mejias, 2019) that both endure and mutate over time (cf. Ladwig et al., 2012; Stoler, 2016).

The coloniality of humanitarian data

Who holds the data? Who holds the power? Data is the power that we have, right? Knowledge is power, it's Foucault, it's basic colonialism, right… this is one of those areas that we need to look at, what does decolonising data mean, right? …you offer communities ways to access and understand data that you’re generating, and, yeah, that's a massive challenge. (Donor representative, I.01.SH.13)

The coloniality of humanitarian data, and the need for its decolonisation, is a debate we already found to be present in our interviews with humanitarian practitioners specialising in data and information management. As indicated in the quote above, data are conceived by this member of the donor community in terms of a form of power associated with ‘colonialism’, which positions those taking data in a hierarchical relationship with communities about whom data are generated. The problem identified in this context is related to community access, as well as to the difficulties of facilitating community understanding of the data that is produced. In other words, the capacity of IDPs is put into doubt here in the very articulation of data as a dimension of the coloniality of humanitarianism. Paternalistic relations of expertise thus underpin this appeal to epistemic justice.

Concerns regarding colonial recurrences in humanitarian data and the challenges of epistemic injustice emerge in various ways in our interviews. For example, another interviewee highlights the problems of ‘supply-driven’ data, or data that is generated by humanitarians for humanitarians:

…a lot of sort of tracking of people is how I often see humanitarian data or needs assessment data … they’re very sort of supply driven, or to an extent. (INGO representative, I.01.SH.25)

Needs assessment data is collected regularly by international organisations involved in the provision of assistance to IDPs in camps. This involves the monitoring of communities within the camps, which facilitates ongoing analysis of the size of displaced populations and the overall level of need. The suggestion by the INGO representative quoted above that this data is ‘supply driven’ indicates that data collection is required by the organisations themselves, rather than by those that they directly ‘serve’. Again, paternalistic relations of expertise underpin this way of thinking, with humanitarians (rather than IDPs) standing as caretakers, who know best how to distribute humanitarian assistance effectively.

While humanitarian organisations do of course ‘serve’ affected communities as part of their mandate, it is important to recognise that their work is orientated to other constituents as well. For example, one NGO representative who we interviewed emphasises the importance of donor accountability to the production of humanitarian data:

…we collect the data because the owners want accountability, the owners want accountability because their constituents are giving them the money and allocating the money, their constituents want their money to be used efficiently so at the end of the day they are really serving their needs … We’re…doing needs assessments after needs assessments …nobody's coordinating them and it's just like incredibly extractive, and that's not accountability, that's just like box ticking at that point, and so… I don’t know where the problem lies but I think at…a certain point, donors have to change the expectations about what data we collect and why. (NGO representative, I.01.SH.16B)

Notable in this statement is that data is clearly perceived as ‘owned’ by data controllers (i.e. donors) rather than by data subjects themselves (i.e. IDPs). Accountability in this regard is focused on donors and – by implication – the national publics whose taxable income funds the activities of donor agencies. The extractivist approach that this generates within the humanitarian field is highlighted as a key problem in this context.

The ways in which humanitarian organisations become caught within this logic of serving the needs of donors to the detriment of affected communities is further referenced in relation to the political economy of the sector and competition between UN agencies, INGOs and donors:

I see competition within the sector driving data hugely. Everybody will have to have be seen as having the best data. It's linked to visibility. It's linked to really who is the leading role. Who is the lead on the ground? …Of course, it's always been [the case in] the NGO world. And it's also in the donor world. Donors want to be seen as leaders. There are countries [that] want to be seen as leaders. …So, we’re all pushing data in different ways that actually support our own objectives. And I think that's driven a certain amount of the data. I do think also technology has definitely helped us a lot. We have better means of getting quick data now. We have a lot more opportunities to get data quickly. And whoever gets [data] out first has the upper hand in terms of getting more funding, … [and gaining a leading] reputation. (Network representative, I.01.SH.17B)

The above quote starkly highlights how data can be a mechanism for economic opportunism as well as for the reputational benefit of powerful donors and agencies. Epistemic injustices come together here with structural injustices, as data is extracted from the communities who are supposed to be ‘served’ by humanitarians, effectively advancing existing structures of domination (cf. Mudler, 2020).

The extractive dimensions of data-driven humanitarianism arise in various ways across our interviews, facilitating further understanding of the coloniality of data-driven humanitarianism and the forms of epistemic injustice within which this is grounded. This is further elaborated by an NGO representative who we interviewed, in terms of the dispossession of agency:

…data is used to dispossess the agency of the different populations… I’m going to sit in my office with my directors and decide what these people need. And basically, I’m using the data but I’m actually removing agency [of the affected populations] from developing. And this is really dangerous …during COVID, you will have seen that more because you cannot go face to face and that's the perfect occasion to stop talking to people and just sit in our offices and discuss between ourselves what these people actually need. (NGO representative, I.01.SH.04)

Indeed, we found numerous references in our interviews to this extractivist drive:

…there is this sort of desire for endless data collection and just endless like extractive processes of how much can we learn about a humanitarian situation and then where all that data goes and to what extent it gets used sometimes is a bit hard to understand. (INGO representative, I.01.SH.25)

…while there's lip service to the affected population generating their own data and sort of a couple of examples of that, really this information has been hoovered up from affected people and [has] been given to people outside of the affected community who are making decisions on their behalf. (Think tank representative, I.01.SH.08)

An extractive approach to data collection and use in this regard is integral to the coloniality of data-driven humanitarianism, with epistemic injustices manifest in the oversight of beneficiaries of protection and assistance as having capacities as ‘knowers’ or as subjects of knowledge.

Extraction is not only viewed as problematic for beneficiaries, but also for the local data collectors who are often central to the functioning of data-driven humanitarianism. The interviewee cited above goes further:

There is a labour rights issue in the way that data is collected. And it's usually collected, in a temporary way, by short term volunteers who are not paid very well and tend not to have benefits… So, that's another point of extractiveness. …when local data collection partners are used as just sort of data hoovers rather than as equal partners with the big donors. (Think tank representative, I.01.SH.08)

Data gathering is viewed here not as a neutral process, but as one involving the extraction of a valuable resource to generate finances and assert leadership. Paternalism is central to this process of professionalisation, as another interviewee indicates:

Is it a task between a load of experts who are having a fight about which kind of paternalistic value they’re going to drop down? (Network representative, I.01.SH.17A)

Concerns about double standards within data practices between the global North and global South and the paternalism of assumptions that humanitarian data is automatically a ‘good’ is problematised by yet another of our interviewees:

I think we have a tendency to gloss over the protection risks for people who are in these countries where we’re providing assistance. Like somehow, you know, because they’re not doing online shopping the way we are and they don’t have bank accounts with Citibank that they don’t need as much, you know, protection of their identity or internet security or having their information out there, and very detailed personal information. And, you know, I think that's naïve on our part and condescending and so paternalistic as well. And then, you know, we’re collecting it and for their good. …we’re so focused on collecting so much and at such a granular level, but is it really in people's best interest to have that information floating around? (UN agency representative, I.01.SH.01)

In this context, the assumption that actions of the humanitarian community are benign and altruistic is directly challenged on the grounds of coloniality. As an INGO representative we cited earlier further explains:

…the way that data is being used now is still quite colonial I would say in nature and still very much like what can we learn about these people and…how do we therefore make decisions the way that we want to make them about how we’re going to address their problems … I think that there's this idea that humanitarians are operating under huge stress and that they’re kind of doing the best that they can and that it's ultimately an altruistic pursuit. (INGO representative, I.01.SH.25)

Humanitarian data experts in this regard explicitly point to the ways in which paternalistic assumptions associated with the coloniality of humanitarianism are further perpetuated through the ‘data revolution’. In the next section, we analyse IDP and stakeholder responses to humanitarian data collection and use in north-eastern Nigeria and South Sudan, to highlight the frictions surrounding epistemic injustices that are generated in an increasingly digitalised context.

Humanitarian, data and ethical frictions

Our interviews with IDPs and stakeholders bring to light a range of frictions that are generated by the coloniality of humanitarian practices in north-eastern Nigeria and South Sudan. We approach moments of friction over the production and use of humanitarian data as important ‘because they influence what is made visible to and knowable by who, and therefore impact profoundly on the development of future knowledge and social relations’ (Bates, 2017: 423). The frictions that we identify in this section relate to which groups are prioritised in the provision of humanitarian assistance; over processes of humanitarian data collection; and over the ethics of data-driven humanitarianism. It is in relation to the latter specifically that we suggest the ‘data revolution’ generates intensified frictions over coloniality.

Humanitarian frictions

The situation for IDPs in both north-eastern Nigeria and South Sudan was particularly difficult during the period when our research was carried out, and not only due to the impacts of the COVID pandemic. There were also ongoing uncertainties over the peace process in South Sudan as well as continued tensions and conflict in north-eastern Nigeria. While the focus of this article is on data-driven humanitarianism, the ongoing recurrences of coloniality in each of these contexts are also evident in the presence of peacekeeping, development and other international agencies and corporations in both regions. Those we interviewed faced instability and uncertainty on various fronts, including regarding their immediate living situation. There had been recent changes to the management of the camps in South Sudan and there were closures of some of the camps in north-eastern Nigeria soon after we completed our interviews. IDPs in both regions faced shortages of basic services, food and water. In South Sudan, there had also been severe floods perpetuating the problems people were experiencing, with many struggling to survive on reduced food rations (see also OCHA 2021a, 2021b).

In this context, several stakeholders highlighted the need to focus support on ‘vulnerable’ groups such as women, children, people with disabilities and elders. This is described both in terms of addressing specific protection needs, as well as providing general assistance to meet basic needs:

Honestly, everyone to be sincere needs to be protected. But women and children need protection to preserve their dignity. They need to be more protected. Children need protection and women are more vulnerable, so they need protection. (State representative, Nigeria, N.01.SH.07)

Most of the vulnerable ones were old men and children, but you cannot determine the adults because they can go out and find something for themselves. (Youth leader, Nigeria. N.01.SH.06)

…when you talk about women's protection and empowerment, you can call it women's emancipation, to eradicate habits that oppress women. …That is what the protection focuses on. For general protection, …we care about each and every IDP in the camp – from elders, children, women, the disabled and other people living with disabilities, etc. and we basically follow a guideline from UNHCR, the 11 codes of persons with specific needs. (International organisation, South Sudan, SS.01.SH.02)

Taken together, these quotes indicate that the protection needs of women and children are identified by many stakeholders as a key priority, along with people with disabilities, health issues and the elderly (see also OCHA, 2021a, 2021b). International organisations do follow distinct criteria in identifying vulnerabilities, although the role of data in identifying vulnerable groups is not always clear-cut.

So how do IDPs view the provision of protection and assistance, and on what basis or set of assumptions? In Nigeria many of those with whom we spoke highlight inequalities in the provision of assistance:

We are not treated equally, tell anybody. Yes, they select small girls, and give them, not old women, like us. Sometimes there is no equal treatment. (Female IDP, Nigeria, N.01.IDP.06)

There is no equal treatment, but they only consider the vulnerability of the person and lactating mothers. (Male IDP, Nigeria, N.01.IDP.09)

There is no equal treatment in any camp, because of the nature of the influx of IDPs, new arrivals to the camp are not enjoying some incentives like shelters, and food. (Female IDP, Nigeria, N.01.IDP.02)

…food was not distributed equally, because they are not distributing it according to tents. (Male IDP, Nigeria, N.01.IDP.09)

During the first phase of our fieldwork in South Sudan, concerns about the generalised lack of food were particular stark, though concerns were also raised about differentiated levels of assistance:

This information, you know they take from those who are orphans and those who are disabled and the vulnerable. I know an orphan who is related to me and he completed a form and was assisted. For me I join the line and get my tent and also get my food. (Male IDP, South Sudan, SS.01.IDP.03)

…we living in the camp, our idea is that if we were provided with adequate assistance we would go to school and get work and there would be no instances of the youth engaging in fight and theft. What brings a situation of theft is because what people in the [camp] have is not enough and are hungry. … If there is adequate assistance all of us would want to go to school and work and not engage in conflicts. If one has a child, one cannot even enrol the child into school. This is what I have to say. (Male IDP, South Sudan, SS.01. IDP.10)

In both north-eastern Nigeria and South Sudan frictions over the provision of humanitarian assistance belie more far-reaching disagreements about who counts as ‘vulnerable’ and on what basis assistance should be provided. Yet, more than simply evidence of disagreement over the unequal provision of assistance, the quotes above suggest that IDP knowledge is effectively subjugated within the humanitarian field. This raises important questions about how humanitarian data is produced and the frictions to which data collection processes give rise.

Data frictions

Beyond frictions over the provision of assistance, frictions also emerge around processes of data collection in both contexts whereby we undertook fieldwork. For example, in South Sudan a male IDP suggests that the ways in which information is gathered effectively excludes men:

I have never spoken to any organisations since the time I was displaced in 2013. However, it may seem that the organisations would speak with women because they are the ones at home and we as men do not stay at home. (Male IDP, South Sudan. SS.01.IDP.09)

This quote identifies a direct example of the ways in which data collection processes can generate gender bias. Indeed, our analysis suggests that other forms of bias may also be embedded within humanitarian data. For example, in Nigeria vulnerabilities are sometimes described by local stakeholders as visible or easy to identify. When asked how to identify a vulnerable IDP, one stakeholder responds: ‘By mode of their dressing�� (State representative, Nigeria, N.01.SH.09). Another suggests that you can know IDPs who are vulnerable by ‘merely looking and through there you can identify them their appearances and places’ (State representative, Nigeria, N.01.SH.05). This is further elaborated by the first stakeholder, who says that: ‘We do identify them by way of their interaction with people, way of talking with people, association and others, mode of their dressing and others’ (State representative, Nigeria, 01.SH.09). Questions arise here as to how biases structure the production of data where vulnerabilities are viewed as obviously apparent and less visible issues risk being overlooked (e.g. mental health, sexuality). Questions also arise about how such processes reinscribe gendered, racialised and ableist conceptions of subjects who have the capacity to know and make decisions (see also Medina, 2013).

For IDPs in South Sudan, a wider culture of insecurity related to long-standing internal conflict has a notable impact on the process of data collection (e.g. see Hutchinson and Pendle, 2015; Read, 2018). For example, one man expressed fear about people coming to collect data, stressing that it is neither clear who is coming nor what is done with IDP data:

…some organisations can come and take information and do not return and sometimes people can come and take the information and we do not know what they will do with the information and sometimes we are afraid. Like I can see that you are a woman and you are taking this information and I do not know what you can do with this. You know that is my concern. (Male IDP, South Sudan, SS.01.IDP.01)

Privacy concerns and the importance of rights-based data justice are of course important here, particularly in situations marked by conflict. Yet our interviews with IDPs in both north-eastern Nigeria and South Sudan suggest that we also need to go further in our analysis of the frictions that emerge around the collection of data, given the more generalised perception of IDPs that information is provided without any response on the part of those collecting data. As a woman in Nigeria describes:

They used to collect peoples account number, not any paper to sign, and they have collected many peoples account number and there is no response. [We] do not hear anything… They are telling lies. (Female IDP, Nigeria, N.01.IDP.06)

This quote raises important questions about epistemic injustice and broader ethical frictions to which processes of data-driven humanitarian assistance give rise.

Ethical frictions

Thus far we have shown how data-based humanitarianism generates frictions both over the provision of assistance as well as over processes of data collection through which assistance needs are identified. Yet it is in relation to ethical frictions – particularly the participation of affected communities and issues of accountability – that we can begin to unpack how the ‘data revolution’ intensifies both epistemic injustices and frictions over the coloniality of humanitarianism. The issue of participation is raised directly by some of the stakeholders in Nigeria, most notably those who are more attuned to the invisible vulnerabilities that we noted in the last sub-section:

Some are identified physically, for example, people with disability, pregnant women, lactating mothers. You can easily identify them. Other vulnerabilities are difficult to identify. But we are working with community hand in hand, and sometimes by our monitoring we can…identify a person who is vulnerable. (State representative, Nigeria.01.SH.07)

What happens is that you cannot identify them through appearance or clothes. You only get the sources through their community leaders. (State representative, Nigeria.01.SH.08)

There are of course a range of issues arising here with regard to community leadership and the potential for participation to be enacted in gendered and ableist terms. Despite this, such quotes suggest that working in a participatory way with local communities is crucial for addressing complex and cumulative forms of need. Going further, we would also suggest that such work is important in addressing issues of epistemic injustice. This requires that attention is paid to how the ‘data revolution’ and its perpetuation of paternalistic expertise works against such efforts.

If we turn to key ethical issues such as informed consent, we start to see how issues of epistemic injustice structure the humanitarian field more broadly. In our research, we found that most IDPs expressed understanding of the questions asked of them by data collectors and that translation was generally seen to be effective. Practices of informed consent appeared to be variable, though on the whole concerns were not raised directly about consent: in Nigeria IDPs suggested that consent is inconsistently recorded (Fayehun and Àkànle, 2022), while in South Sudan several reported that they were not asked for consent prior to sharing their data (Logo and Jones, 2022). Regardless, what we find striking is the more generalised frustration on the part of IDPs regarding the use and efficacy of data. As a man in Nigeria suggests: ‘They will just collect data and go. Our people here have participated in many interviews, but nothing has changed’ (Male IDP, Nigeria, N.01.IDP.09).

Indeed, a lack of return visits by those taking information was highlighted in both Nigeria and South Sudan as key concerns, along with the failure to listen and respond to the information that is provided. As one woman in South Sudan expresses:

They do ask, but it is not the way you are asking. They come to us at home and they ask us about our needs. They interviewed me and I explained to them that I need the plastic sheeting, malaria drugs and the needs of a household. (Female IDP, South Sudan, SS.01.IDP.06)

These quotes point to the ways in which data collection processes often fail to meet IDP expectations, generating epistemological closures that privilege some forms of knowledge over others, while perpetuating extractivist processes of data collection. So how does the ‘data revolution’ and the paternalism that this exacerbates further generate epistemic injustice in the humanitarian sector? In the final section, we will return to our interviews with humanitarians to reflect on the ways in which the digital turn has further perpetuated paternalistic forms of expertise in terms that rest on a limited conception of accountability.

The limits of accountability

The accountability of humanitarian and donor agencies to affected communities has been a growing area of concern over recent years, as is evident in the Grand Bargain of 2016. ⁵ Nevertheless, the limits of accountability and participation come out strongly in our findings from interviews with humanitarian data experts. For example, a donor representative suggests:

…we’re really leaping miles ahead with all of the data actors but we’re not really leaping miles ahead in terms of community engagement and community trust and the accountability side of that work, right? Participation revolution, well, it's not revolution because it hasn’t happened, right? (Donor representative, I.01.SH.13)

As indicated earlier, this is perhaps unsurprising given that humanitarian data is largely oriented towards benefiting humanitarians:

…almost all humanitarian data is structured to benefit humanitarians, and we probably claim, okay, our programming is then therefore benefitting affected people and that's how we justify the focus, but at the end of the day it's data for us and for our purposes …it's an economy that is built around this data …politics and governance built around those data yet effected people don’t really see tangible results of it, so I think one thing…is trying to make data benefit affected people more directly. (NGO representative, I.01.SH.16B)

Another of our interviewees suggests the need for a review of existing models of participation and community engagement in the collection and use of data:

…they do a needs assessment and they call that engagement or, you know, they do a complaints box but they put it, you know, in the corner somewhere where it's really hard to reach … I think that there can sometimes be this view that, you know, if we include these individuals in data collection or collect more data on them that we’ve done the job of including them in programme design and I think, you know, we need to actually move towards much more participatory formats for that and different ethical approaches to how we collect and use that data. (Network representative, I.01.SH.05)

Reflecting back on the supply-driven conception of data discussed earlier in this article, what this quote points to is the way in which affected communities are largely side-lined within practices of data-driven humanitarianism.

In this context, a rights-based model of data justice is deemed crucial by many humanitarian data experts. For example, one highlights how ‘advocacy towards data is not going hand in hand with the advocacy towards legal frameworks on data’ (UN agency representative, I.01.SH.28). Another stresses how a rights-based approach requires that affected communities ‘claim their rights… and maybe make claims against humanitarians or against government’ (Think tank representative, I.01.SH.08). A concern with accountability again arises here:

I think if we take more of a rights-based approach and understand that it's not a small thing when you’re asking for people's data, even if it's not biometric but especially biometric data, you know, it's really not a small thing at all. It's a huge thing that you are taking from them, it's almost a transaction rather than you, you know, necessarily providing them with a service that's more beneficial to them than they are to you. So I think that having a stronger rights-based approach to the way that data is collected and used is really important, and then I think we need to be accountable for failing – I think we need to have more transparency and accountability for how data is used or not used and people should be justifying, you know, the reason to collect new data rather than assuming that it's always going to be beneficial. (Network representative, I.01.SH.05)

What emerges here is a sense that accountability is firmly tied here to a rights-based approach to data justice, which will enable affected communities the opportunity to hold those collecting and using their data responsible for any failure.

Indeed, data ownership is widely seen as an important issue in connection with accountability and data rights by our interviewees, including with regard to the collection, access, use and control of data:

…with the concept of localisation of aid and accountability to affected populations, I think there's actually under-investment and there hasn’t been enough talk around how populations – how people affected – can actually hold the humanitarian system accountable through control of their own datasets, and that's sort of an emerging thing that we’re really interested in. (INGO representative, I.01.SH.10)

…we need to be asking the people that are actually affected and they should be the ones to determine what data… not like what data's being collected but they should be actually involved in that collection effort and have some way to actually own their data, have rights to see it, control it, know how it's been used, and we’re pretty far from achieving that I think. (INGO representative, I.01.SH.10)

These quotes suggest that there is significant support within the field of data-driven humanitarianism for a rights-based approach, which recognises the need for affected communities to maintain ownership over the production and use of such data.

Yet beyond the question of rights or of the procedures required to ensure the just production and use of humanitarian data, we want to go further to emphasise how questions of accountability relate to deeper-rooted questions of epistemic justice. Indeed, our analysis suggests that some humanitarian data experts are themselves working to challenge the limits of an approach to accountability based on rights alone:

I think we should be accountable… It goes back to the people we collected information from …I think that transparency and accountability to people should apply also to the data discussions… I mean, information is power. I think for communities to understand what is being said about them and their needs and their situation this…should be an open thing. (UN agency representative, I.01.SH.23)

I think access is a big thing, right. I mean, access for people like access to information, but also access to voice, information, and opinions. (UN agency representative, I.01.SH.23)

This emphasis on ‘voice, information and opinions’ is significant, because it points to the need for the revaluation of IDP knowledge and the importance of respecting displaced persons as subjects with the capacity to know and make decisions.

It is worth reflecting here on the ways in which a logic of extraction is not only perpetuated by the increased value attributed to data within the humanitarian field, but also is integral to the paternalistic ethical framework on which such data relies. As one of our interviewees suggests, the communicative model on which humanitarian accountability rests is ‘this sort of very controlled conversation’ (INGO representative, I.01.SH.25). The importance of a more dialogical approach to communication is proposed as an alternative in this regard:

… how could you better sort of work with communities to say you also define the indicators, so in the context of your life like an aid response is one facet of what you need to sort of recover or thrive, and so what are your priorities, and not just your priority needs but just your priorities in life and how much of those are you able to sort of meet or address. (INGO representative, I.01.SH.25)

There is much to unpack in this statement, but what we want to emphasise is that advocating for a model of communication based on ‘listening’ can be interpreted as a commitment to epistemic justice:

… the whole system is not structured to be listening … we will now go back to communities and say this is what we found, this is how we analysed it, this is what humanitarians have said, does this still resonate with you, are you able to use this for your own purposes based on what this data says, what are your recommendations … then taking that again back to sort of humanitarian coordinators, and now we’ve added a final step which is sort of how then do we turn that into more of a public good, so if there's a community that sort of wants to use that for its own advocacy, how do we make sure that they have not just kind of this is what we did and this is what we took and here's what we did with it, but more here's all the data and how do you want to use it. (INGO representative, I.01.SH.25)

Striking here is the way in which this emphasis on listening resonates with the statements of IDPs analysed in the last section, who express frustration with the ways in which data is collected without any impact for those from whom it is extracted. In recognising IDPs here as subjects with the capacity to know and make decisions, the superiority of those deemed as having a duty to care is undermined.

Indeed, it is perhaps not surprising in this regard that the interviewee cited above goes on to explain: ‘we get a lot of pushback because people (i.e. humanitarians) just feel like what we’re telling them to do is really going against the whole aid machinery’ (INGO representative, I.01.SH.25). Elaborating further, they say:

… a lot of the push back that we get is that people don’t feel that they can act on community feedback… when there's a needs assessment and the needs assessment has been devised by taking cluster services and deciding what questions you would need to ask to determine which of those services to deploy and then just like stacking those questions on top of each other, the understanding of need that you get is very driven to the response that's probably already going to happen… so then the feedback that we get you know, it's not always negative and sometimes it says you know, that some things are going fine, but most of the pushback that we get is well of course we can’t act on that [feedback from beneficiaries]. (INGO representative, I.01.SH.25)

Listening in itself becomes problematic in this context, precisely on the grounds of the duty of care:

…if someone's never gotten anything then we start asking them all these questions and then maybe they still don’t get anything… we sort of don’t want to do harm in that sense. (INGO representative, I.01.SH.25)

To put it another way, ‘listening’ to affected communities – along with their expectations that the provision of data will lead to their needs being met – is in effect precluded in this statement on ethical grounds. ⁶ In other words, epistemic justice or not listening stands as the very grounds on which a humanitarian ethic of ‘do no harm’ relies. This not only points to the recurring coloniality of humanitarianism, but also to the perpetuation of colonial dynamics through a ‘data revolution’ that intensifies paternalism through professionalisation and that extends humanitarian expertise at the expense of the experiential expertise of displaced persons.

Conclusion

What does the humanitarian ‘data revolution’ mean for people who are displaced to camps across north-eastern Nigeria and South Sudan? In this article, we have highlighted the ways in which the digital turn perpetuates epistemic injustices through the extraction of humanitarian data from IDPs. The analysis has shown how IDPs are systematically disregarded as subjects of knowledge with the capacity to know and act on the data generated about them. Although humanitarian experts are increasingly aware of the extractive practices within which the coloniality of data is grounded, a range of frictions nevertheless emerge surrounding humanitarian provisions, data collection and the ethics of data-driven humanitarian assistance. We argue that paying attention to these frictions is important, because it enables reflection on what is made visible and knowable to who, while opening alternative perspectives to dominant forms of knowledge. In confronting the coloniality of data and the frictions that this generates, humanitarian data experts point to the importance of data rights, donor and agency accountability and the ownership of data by beneficiaries. Some also emphasise the importance of actively listening to the demands of affected communities. Nevertheless, we have suggested that humanitarian ethics of ‘do no harm’ remains a key sticking point here, with a ‘duty of care’ in contexts of data proliferation effectively working against moves towards epistemic justice. Data-driven humanitarianism in this regard serves to condition the intensification of paternalistic dynamics, and as such is characterised by recurring colonial dynamics as well as intensified frictions that bring epistemic injustices into sharper focus.