Introduction

Humanitarian efforts to foster digital work for refugees is a rising phenomenon. Ranging from basic digital literacy courses to long-term ICT trainings and tailored support to obtain online consulting work, often through digital work platforms, a range of international and national actors are working in this area (Easton-Calabria and Hackl, in this issue). Alongside humanitarian organisations such as the World Food Programme and Norwegian Refugee Council are development actors like the International Trade Centre and the International Labour Organization, and even private sector actors and supporters like Upwork and Tent (Upwork, 2022).

Non-governmental organisations (NGOs) working to place refugees in remote digital work today find themselves in the peculiar position of acting as online market intermediaries between refugees and corporations in the digital economy. This ranges from helping refugees sign up to online job platforms to securing work contracts for them directly (UNDP, 2019). Yet if labour exploitation occurs once humanitarians are no longer involved, such as refugees not being paid for their work, or if refugees try but fail to secure online work after trainings, do the organisations that involved them in the digital economy bear any responsibility?

There is little explicit scholarship on the concept of ‘digital responsibility’ in humanitarianism or specifically related to refugees, let alone related to humanitarians seeking to link refugees to the digital economy. Existing definitions of digital responsibility often centre on both accountability and liability: humanitarian actors should be held accountable and liable for actions that go against the so-called humanitarian imperative to do no harm. However, some research notes accountability gaps in relation to humanitarians and their beneficiaries, including refugees (Martin et al., 2021): instead of humanitarians bearing responsibility for the outcomes of humanitarian assistance on refugees, it is instead ‘upward accountability’ to donors that is the primary mode of accountability engagement (Daun, 2020).

Yet outcomes of digital work for refugees are particularly important to consider, as this type of work is generally linked to the global gig economy, itself inherently intertwined with job precarity and global inequality. While this has been acknowledged by many humanitarians I interviewed for research on digital work for displaced people between 2019 and 2021,¹ there does not appear to have been organisational deliberation about the ethics and potential harms of this premise. However, work on this broader area exists (Jones, 2021), and critiques are likely to grow. Some notable organisations such as Humans in the Loop² and DignifAI³ provide ongoing support to refugees seeking work in the digital economy, yet refugees generally find themselves post-training competing alone for small-scale gig contracts. Risks lie in a range of areas, including indecent work, risks to health and safety, scams and fraud, and possible debt accrued through unpaid time or buying equipment for digital work that does not appear (Rushworth and Hackl, 2022).

At the very least, organisations offering digital trainings to refugees should engage in robust monitoring and evaluation (M&E): long-term tracking of refugees’ livelihoods outcomes post-training, adjusting programmes based on these outcomes and being honest about whether these outcomes justify the existence of these trainings at all – not to mention the time, effort and hope that refugees put into them. Part of this tracking should entail a deep understanding of the financial outcomes of ‘gigs’ that refugees secure, as well as a critical consideration of the type of work that refugees undertake. Based on my interviews with humanitarian organisations, this does not occur on a widespread basis.

Further, many interview informants concede that the entire model of mobile and flexible digital work for refugees in low-resource contexts is dependent on the competitiveness of refugees in poor countries as a cheap labour force. Much online work is extremely low-paid, such as that which consists of tedious hours of clicking through images, ultimately to train the machines that the world’s largest corporations (think Amazon) develop for outrageous profits (Webster, 2016). As one head of a major digital livelihoods skills training programme for refugees in the Middle East, sub-Saharan Africa and South America told me, ‘The work is high volume, low skills … we’re not prepping people to become software developers.’ Alongside the digital divide there exists a further divide between the work that many humanitarians expect for themselves compared to that which they are prepared to offer refugees. Be it employees in Geneva or staff working directly with refugees in the so-called Global South, humanitarians expect decent salaries and steady contracts; hopefully they also expect this work to be stimulating and offer opportunities for upwards professional mobility. However, these and other components of decent work are rarely available to refugees entering the digital economy. Digital technology has been espoused as a tool for ‘self-reliance and positive change’ (UNHCR, 2022) and even a ‘saving grace for refugees’ (Koigi, 2022), while the reality is of course much more complex; why should the humanitarian community not be more aspirational about the type of work (online and offline) that they support refugees to undertake?

And, if it turns out that refugees on a large scale are in fact unable to obtain digital work despite skills trainings by NGOs, then what purpose are these trainings actually serving? The possibility that these programmes perpetuate restrictive host country policies surrounding refugees’ (lack of) right to work, or endorse donors’ exit strategies rather than offer refugees protection and dignity, as current rhetoric on digital inclusion suggests (e.g. UNHCR, 2022), are concerns with which organisations must further reckon. Allaying these concerns requires robust M&E, as mentioned above, including explorations of whether refugees are pursuing digital work as a first or last resort. This in turn could shed light on other areas for action ‘in real life’.

If one takes as a starting point that moral and ethical norms and considerations should guide human behaviour, then this rationale extends to the creation, use and outcomes of digital technologies, as discussed in relation to ‘ethical AI’ (Floridi et al., 2018) – and relevant for humanitarian digital engagement with refugees and other populations. So-called digital responsibility overlaps with many existing important considerations around refugee rights, including their right to work and at work. However, a more explicit conversation is needed within and across humanitarian organisations as to the level of responsibility humanitarians hold relating to both the intended and unintended outcomes of skilling refugees to join the digital gig economy. While intended consequences can be summarised as providing refugees with access to work and the secondary improved benefits this can bring, such as increased well-being, unintended consequences include but are not limited to introducing refugees to low-paid, low-quality work that does not offer sustainable livelihoods – and even a lack of digital work at all, despite the extensive training and time, and sometimes money, invested (Rushworth and Hackl, 2022).

Ethical engagement in technology is not just about data privacy and use; indeed, it extends much further to a need for humanitarian organisations to develop coherent, contextualised sets of norms and frameworks for responsibility and protection in the digital sphere, including the digital economy. We can look to the public and private sectors to consider how a framework of digital responsibility in this area could or should develop. Lobschat et al. (2021) focus on ‘actionable guidelines’ for corporations in relation to digital responsibility, while similar efforts to develop normative frameworks for responsible digital innovation emphasise responsibility within innovation process and outcome (van de Poel and Sand, 2018).

This project of digital responsibility in relation to refugees and the digital economy is currently lacking but is incredibly important. As part of this, humanitarians must clearly define both what is feasible for refugees to achieve through digital skills training and what ‘success’ in the digital economy looks like. This is all the more crucial because the casualisation of work, the growth of the gig economy and the largely unregulated digital economy all point towards the present and future precarity of digital work in particular (Heeks, 2017). If humanitarian organisations are serious about protecting and upholding refugees’ rights, fostering self-reliance and increasing digital inclusion, then centring discussions on their own digital responsibility is an important place to start.