Volunteerism is widespread in the translation sector, but the practices associated with it and its ethical import have so far received very little critical attention. This article critiques one of the most high profile beneficiaries of volunteer translation, Translators without Borders, which presents itself as a charity but operates as a corporate concern, with a leadership composed primarily of major industry players. TWB adopts an asset-centred, platform-based, top-down model that offers massive scaling possibilities and reflects a corporate vision of the translation community. It provides a clear example of the wider shift from artisanal to industrial to platform economy as it plays out in the translation field. To demonstrate the potential for volunteer translation to be situated within a more solidary and equitable context and provide an example of one possible alternative to the platform-based paradigm, we discuss the practices of another humanitarian NGO, Solidarités International, which runs a paid internship programme and adopts a small-scale, peer-based, horizontal model with a strong focus on early-career translators. We ask who ultimately benefits from the exploitation of free labour and focus on identifying practices that enhance or jeopardize the professionalization and stature of the translators involved. We further discuss how the linguistic assets produced by volunteer translators can generate saleable intellectual property and how this can lead to conflicts of interest and support patterns of inequality in the wider social context.
Translators, whether professionally trained or otherwise, volunteer their time and skills in many contexts, including humanitarian assistance and political activism. Volunteering for humanitarian organizations such as the International Committee of the Red Cross (ICRC) and the United Nations High Commissioner for Refugees (UNHCR) has received very little attention from scholars of translation but is partly addressed in a small number of recent studies. Delgado Luchner and Kherbiche (2018, p. 418) confirm that “the overwhelming majority” of interpreters working “in refugee camps or UNHCR offices” are incentive workers, that is refugees who provide interpreting for very low payment, which may take the form of cash vouchers or in-kind goods.11 Moreno-Rivero (2018) features an interview with a former Senior Project Officer with Translators Without Borders, the subject of the current study. No critical questions are posed and the focus is on research the Officer undertook for TWB and Save the Children during the Greek migration crisis, stressing, in the Officer’s words, that “*t+hrough research TWB recognizes the gaps in the field, measures the effectiveness of certain tools and formats, and acquires evidence, data and statistics to strengthen its advocacy” (ibid., p. 154). Federici et al. (2019, p.6) briefly refer to volunteering for humanitarian organizations in order to stress the importance of investing in technological resources, arguing that “the expansive demands of language access are likely to outstrip the internal resources of most organizations in the humanitarian sector, and only those with the greatest budgetary resources will be able to contract translation services, or materially support key volunteers in the translation domain”.
Volunteering in the context of political activism has attracted more interest and has generally been approached more critically. Among others, Boéri (2008, 2009) offers a detailed, critical analysis of the work of Babels, the international network of volunteer translators and interpreters who cover the linguistic needs of the Social Forums, Baker (2013) examines the political positioning of volunteer translators involved in collectives such as Tlaxcala and Translators Brigade, and Baker (2016) provides a critical account of the ethos and output of volunteer subtitlers who supported the work of two collectives of film makers involved in reporting events during the Egyptian Revolution. Selim (2016) is a first-hand account of the author’s personal experience as a volunteer subtitler for Mosireen, a collective of film makers active during the Egyptian revolution. These studies challenge the prevalent model of treating volunteer translators as service providers by developing alternative discourses and practices that empower translators and recognize their labour – paid or otherwise – as a valuable contribution to society rather than a cheap source of additional profit for the corporate world.
The term volunteer translation overlaps with but is broader than terms such as crowdsourcing, which assumes that the unpaid translation work is undertaken in digital space and is solicited by content owners such as Twitter and Facebook (McDonough Dolmaya, 2020), and user-generated translation, which suggests that those producing the translations are also its ultimate users. We opt for the broader term because despite the increased reliance on digital platforms in soliciting and undertaking unpaid translation work in recent years, most of the ethical issues we discuss are not limited to crowdsourcing in the strict sense. The term ‘volunteer translation’ also allows us to engage specifically with the ethical and social implications of unpaid translation work, whether offered as an act of charity or solidarity with disadvantaged or threatened communities. At the same time, we will be highlighting certain aspects of the widespread practice of crowdsourcing where relevant in order to situate the examples of volunteer translation we discuss within the wider context of the platform economy and the widespread exploitation of digital labour (Morozov 2013; Scholz 2014a).
Volunteer/Crowdsourced Translation and the Platform Economy
Lanier (2013,p.53) warns that “digitizing economy and cultural activity will ultimately shrink the economy while concentrating wealth and power in new ways that are not sustainable”, citing translation as an example. “The act of cloud-based translation”, he explains, “shrinks the economy by pretending the translators who provided the examples do not exist. With each so-called automatic translation, the humans who were the sources of the data are inched away from the world of compensation and employment” (ibid., p. 20). Given the centrality of translation in the information society and the growing interest in its social and political impact, it is important to explore how these general industrial economy trends apply more broadly to the field of translation – once considered as artisanal economy – and how alternatives to a platform-based approach might help counter some of these trends and restore an element of parity to the system. The relevance of this discussion thus extends beyond our immediate target audience of translation scholars, raising issues that are pertinent to social theorists, scholars of political economy and digital culture, as well as non-specialist audiences interested in the ethical questions it raises and the power relations underpinning the political economy of volunteer work.
The new platform economy that has replaced artisanal economies such as those of traditional translation is specifically directed at reducing the value of human labour (Rushkoff 2016,p .19), with technology playing a major role in the process. The impetus to devalue human labour underpins the extension of “the extreme efficiencies of digital networks” to new areas “in such a way that the sources of value, whatever they may be, are left more off-the-books than they used to be” (Lanier 2013, p. 66). Rushkoff (2016, p. 7) adds that corporations introducing new technologies “are free to disrupt almost any industry they choose – journalism, television, music, manufacturing – as long as they don’t disrupt the financial operating system churning beneath it all”.
Crowdsourcing, a practice that is widespread in the field of translation and whose ethics have been rarely questioned, 2and then only from the perspective of its impact on the profession (Baer 2010; McDonough Dolmaya, 2011), is a major feature of the platform economy and a prime example of the devaluation of human labour that it enables. As Rushkoff points out, drawing on Scholz (2014b), “in crowdsourcing there’s no minimum wage, no labor regulation, no governmental jurisdiction” (2016,
p. 50). With a high and increasing proportion of translators working as freelancers, translation lends itself readily to crowdsourcing projects initiated by the likes of Twitter and Facebook. The integration of linguistic assets such as translation memories and the widespread use of word-based rather than hourly- or project-based pricing schemes have accelerated the commoditization of the sector to the point where the platform TM Town, owned and operated by Proz.com33 (a membership-based network website targeting freelance translators), invites freelancers to upload their own resources, including translation memories, in order to improve their ranking in the bidding process on new translation projects, thus shifting the focus from skills to assets, and from value creation to value extraction (Figure 1).
Figure1. Screenshot from TM Town Page ‘10 Reasons to Upload Your Prior Work to TM Town’
Scholars and practitioners of translation have rarely shown awareness of the cynical aspects of such crowdsourcing practices, and so far have never examined the motives behind soliciting volunteer translation for humanitarian organizations. This is not surprising, given that the humanitarian rhetoric is rarely questioned outside those areas of scholarship that are directly concerned with the study of humanitarian and non-governmental organizations. The humanitarian rhetoric has been widely instrumentalized by the US and UK during their various invasions of the Middle East, with implications for NGOs, described by Colin Powell in 2001 as “such a force multiplier for us, such an important part of our combat team” (Krausse, 2014,p.18). The perception of NGOs as implicated in military operations may be one of the reasons why “humanitarian relief is a very reflexive and very self-critical field” (ibid., p.126). The same cannot be said of the field of translation, where this rhetoric continues to be accepted at face value, and where volunteering for humanitarian causes is typically couched in the language of charity rather than solidarity, as evident in some of the quotes from volunteers cited on the Translators Without Borders website, Volunteer section: “I’m well-off in my world. Many others need help in theirs. That’s why I volunteer”; “This work is immensely satisfying – particularly when I can see how I am helping to make a difference”; “The sense that people are genuinely helped by my translation makes me happy” (Figure 2). Construing volunteer translation as an act of charity rather than solidarity has consequences for the level of critical awareness with which we are likely to approach it.
Figure 2: TWB – In the Words of Our Volunteers (accessed 25 October 2019)
Unlike charity, solidarity is reflexive and is able to critique itself (Rai 2018:14). Khasnabish (in press) highlights two further characteristics whose relevance to the current discussion will become clear. Solidarity, he explains, is “a transformative relationship for those involved in forging it, not a thing to be achieved; second, it is grassroots in nature and often constructed from the margins, not something imposed from above”. As such, many acts of solidarity are not sanctioned by mainstream institutions and may attract critical attention. Charity, on the other hand, is rarely treated as a potentially controversial act, and hence is more likely to escape scrutiny. So far, for instance, the limited literature on crowdsourced translation has focused on topics such as motivation (McDonough Dolmaya, 2012; Olohan, 2012, 2014) but has rarely engaged with its ethics or situated it within wider critiques of the platform economy and digital labour. Charity is also not a transformative relationship for those who offer it, as it is typically restricted to helping those in need without expectationof material return. Nevertheless, registered charities are obliged to operate under formal and transparent guidelines to ensure that all donations are properly allocated to the recipient community and stringently managed. This requirement, as we will demonstrate, is absent in the case of some organizations that solicit unpaid translation work as a form of charitable contribution.
Formal non-profit charities have long operated by appealing to governments, corporate donors and the public for donations, which are used to cover expenses and pay suppliers and skilled professionals adequately for their aid services. In some cases, the donation being sought is not financial but actual know-how, often embodied by reusable or codifiable solutions. Seeking such donations by harnessing what Shirky (2010) refers to as the ‘cognitive surplus’ through crowdsourcing has been facilitated by technological developments. With the accelerated consolidation of human knowledge into databases, this leads to a disruption of professional practices (Piróth 2016) and requires engagement with the ethical and social implications of free labour. In addition to its local and discrete charitable effect in the form of delivering translated material, for example, crowdsourcing may also create long-term intangible assets: digital bilingual databases such as translation memories, glossaries and corpora that are of intrinsic value to the translation market. These assets are not localized to the charity recipients: they can be stored, managed, replicated and transferred by their curators. The use of such assets is not regulated by any special regime, and there is no standard for tracking their deployment in other contexts. In short, saleable intellectual property can be generated, posing valid socio-economic questions in our increasingly data-based economy. We demonstrate this issue in section 5 by exploring the close collaboration between the charity Translators Without Borders – the subject of the current research – and the for-profit tech giant Microsoft. We describe how, in a Microsoft-funded project, Translators Without Borders used the unpaid labour of its volunteer translators to produce Swahili language assets that were subsequently integrated into Microsoft’s various commercial tools in 2015.
In what follows, we contrast the practices of two organizations that aim to address humanitarian needs and that have a history of making extensive use of volunteer translators in order to highlight the ethical issues involved in offering free labour to different parties, whether in digital or physical space, and to situate discussions of unpaid translation work within the wider context of the platform economy.
Solidarités International and Translators without Borders
The two organizations we focus on offer two markedly different models of collaborating with volunteer translators. Solidarités International (SI) runs a paid internship programme that adopts a peer-based, horizontal model with a strong focus on early-career translators. Translators without Borders (TWB), on the other hand, adopts an asset-centred, platform-based, top- down model that offers massive scaling possibilities and reflects a corporate vision of the translation community. SI’s internship model operates on a small scale and aims to integrate translators with the rest of the organization’s staff. Linguistic assets created by translators within the internship framework are managed in close collaboration with SI. On the other hand, TWB aims to centralize outsourced translation tasks from many non-profits and to complete them using a free crowdsourcing model through a scalable platform suitable for hundreds of potential NGO clients and thousands of volunteer translators. Linguistic assets created through the platform are managed by TWB’s leadership. An emerging strand of social movement studies has begun to engage specifically with the effectiveness and positioning of volunteer translators in these two markedly different contexts – grassroots vs. top-down models of organization (Doerr, 2018) – but much more still needs to be done.
Our account of Solidarités International’s internship programme is informed by first hand involvement of the first author (Piróth) with the organization as a volunteer, initially in relation to terminology coordination and project management, and later in training and tutoring. Our account of Translators Without Borders practices, on the other hand, is undertaken from an external point of view, building on earlier critique discussed in Baker (2006, 2010). We consider this two-pronged approach fit for purpose in this case, since our primary aim is not to perform a point- by-point comparison but to explore whether probity, transparency and conflicts of interest meet the usual requirements imposed by established charities, and to identify key issues that impact the professionalization and stature of translators in various collaborative setups. Our critical analysis of TWB practices and ethos could not be sensibly based on data collected through interviews with the organization’s leadership. As scholars such as Berezin (2007, p. 132) have noted, it is uncomfortable and unproductive “to interact directly with groups whose views one not only does not share but for whom one feels a certain distance—and distaste”. More importantly, Berezin’s comments on the meaninglessness of conducting interviews with
members of France’s right wing National Front party in the context of her ethnographic study apply equally to our critical study of TWB (ibid.):
Interviewing party operatives as to their strategies and ideas has limited value. No party strategist of any ideological persuasion is likely to provide even the most innocuous looking of academics with their game plan. Nor are these operatives likely to provide any political ideas that are not already covered repeatedly in diverse party documents.
We have therefore opted to draw instead on publicly available data such as TWB’s own website, discussions on platforms such as Proz.com in which TWB representatives and volunteer translators have been involved, TWB declarations to the Internal Revenue Service as a tax exempt organization, and published records of talks by TWB executives.
Founded in 1980 and based in Clichy, near Paris, Solidarités International (SI) currently operates aid programmes in eighteen countries devastated by political conflicts, epidemics and natural disasters. With almost 2000 national and international staff, it helps around four million people worldwide. In recent years, its annual budget has been around 70- 80 million euros, with over 90% consistently allocated to its humanitarian programmes to assist populations in need. Its publicly available annual reports quote the following figures: 4
2013: 70.33 million €, 93.6%, over 5.8 million people helped;
2014: 72.5 million €, 93%, more than 5 million people helped;
2015: 69 million €, 91.3%, 3.8 million people helped;
2016: 71 million €, 91.5%, almost 4 million people helped;
2017: 79 million €, 92.1%, nearly 4 million people helped;
2018: 86 million €, 91.4%, nearly 4 million people helped.
The organization publishes its accounts transparently and undergoes external audits regularly to ensure and demonstrate the proper use of resources. It is among the 91 French organizations that hold the Don en confiance (Donate in Confidence) accreditation (Don en confiance 2018), which requires NGOs to adopt stringent measures regarding transparency, efficiency and potential conflicts of interest. The approval of a dedicated independent organization that lists a French ministry among its partners, and another among its supporters, helps reinforce donors’ confidence that SI’s aid programmes and avowed vision are supported by a robust internal structure.SI has over 50,000 active donors, including major international financial backers, and its key communication materials have to be available in French and English. It does not employ full-time in-house translators, since the demand for translation is insufficient and variable. There is therefore a recurrent need for external French-to-English translation. Other translation needs, to and from the languages used in the countries where the NGO operates, are usually handled locally by SI’s national staff. In terms of organizational structure and governance, full membership of SI can be obtained solely by first doing fieldwork or by spending years in a logistics or administrative role. Only full members are eligible to join the board and must first disclose any potential conflicts of interest. These requirements are standard practice for humanitarian NGOs. For example, Doctors without Borders (Médecins Sans Frontières – MSF) applies the same approach, as a colleague was informed during a telephone call to MSF’s office in Sydney (Vivian Stevenson, personal communication, November 2014). In response to a specific question about possible exceptions for high-profile individuals, he was told by an assistant to MSF Australia’s Board that even Bill Gates would not get an ordinary – let alone a board – membership through monetary donations: he would have to toil at the coalface first. Or as the MSF representative put it, “you earn your stripes”.
The non-profit Translators without Borders (TWB) is often mentioned whenever translation for humanitarian causes is discussed, and its agenda tends to be embraced uncritically by scholars of translation. The Dublin City University led INTERACT project (INTERnAtional network on Crisis Translation), for instance, lists TWB as a partner organization. Although its name may suggest otherwise, TWB is not a front-line humanitarian NGO but an independent non-profit providing linguistic support to humanitarian and other organizations. Launched as Traducteurs sans Frontières in Paris as an offshoot of the for-profit translation company Eurotexte, the organization started to provide free translation to select NGOs in the mid-1990s, by donating the work of Eurotexte’s paid, in-house translators. The conflict between humanitarian and commercial agendas and the resulting narrative incoherence were discussed over a decade ago by Baker (2006, p. 157-162). TWB later adopted a large-scale crowdsourcing approach that extended its pool of translators well beyond its in-house team, imposed professional credentials as entry criteria and offered zero payment to volunteers. It has attracted thousands of freelance contributors over the years through extensive marketing and PR, and has become a household name. As of this writing, TWB has donated over 82 million words (Figure 3) across all languages (over 190 language pairs)
Figure 3: Words Translated Counter on TWB’s Homepage (accessed 25 October 2019)
Unlike MSF or SI, TWB has a strong top-down corporate structure, as discussed in more detail below. According to TWB’s own IRS 990 declarations (TWB 990, 2015, TWB 990, 2016), the organization has no written conflict-of-interest policy to date. TWB thus cannot meet the requirements of Don en Confiance and similar independent bodies. Nonetheless, TWB has received high-level recognition: for instance, in October 2017, in her contribution to the House of Lords debate on Sierra Leone, Baroness Coussins (Vice President of the Chartered Institute of Linguists in the UK) drew attention to the important role played by TWB volunteer translators in assisting recovery from Ebola in the region (Coussins 2017). Just a year before, in October 2016, TWB’s chair Andrew Bredenkamp was the invited keynote speaker at the European Commission’s Translating Europe Forum in Brussels. He was warmly welcomed by Kristalina Georgieva, who had just resigned from her position as the Commission’s Vice President to become the CEO of the World Bank. Mr. Bredenkamp shared information on TWB’s involvement in the Ebola crisis and TWB’s work in Haiti after Hurricane Matthew in 2015. We take a closer look at these flagship projects later in the article, as we examine a range of issues that may serve as yardsticks by which to critique the ethical practices of TWB, using SI as a potential alternative model for organizing volunteer translation work for humanitarian purposes.
These issues include the translators’ status and ethos; the limits of unpaid work; the need to trace the money trail to establish who ultimately benefits from volunteer work; and the path of evolution followed by putative non- profits such as TWB as opposed to that pursued by bona fide humanitarian organizations.
Translators’ Status and Ethos
As the high-level appreciation received in the House of Lords and in the European Commission indicates, Translators Without Borders successfully raised awareness of the importance of language and translation; Federici et al. (2019, p. 5) make this point explicitly. But what of translators themselves? To answer this question, it is worth looking more closely at how TWB addresses different audiences.
When targeting humanitarian organizations and the general public, the organization emphasizes the vital importance of translation; however, when targeting volunteer translators, TWB depicts translation in the FAQ section of its Workspace as an unbudgeted afterthought (Figure 4):
Translation is thus presented as vital or incidental, depending on the audience. The dissonance between the two stances is worrisome in light of the emerging employment precariat in society at large. Indeed, as one scholar of translation notes, “The low status that translators are associated with stands in contrast to the volume of translation work that is carried out worldwide, which has increased under the influence of globalization” (Tesseur 2014, p. 31). Hence, while translation may be doing fine, translators apparently are not. This inversion of benefit is a familiar
historical theme and an inherent feature of the platform economy, but there is no compelling reason why it should be accepted at face value. Demonetization, commoditization and deprofessionalization are unlikely to boost the net worth of society’s cognitive capital, whether in the field of translation or in other areas of the economy.
SI adopts a different approach to volunteer translation that contrasts markedly with TWB’s practices. It set up an external pro bono network of freelance translators in 2007,5 with a dual aim: to provide free linguistic aid to the organization and to create a workspace where qualified translators (including career starters) could collaborate, network and develop their skills. Since previous experience was not a prerequisite, the team was, from the outset, a mix of qualified professionals at the beginning of their career and experienced colleagues. The possibility of working with senior colleagues and receiving detailed feedback turned out to be particularly attractive to young colleagues, who felt that it accelerated their transition from qualified but inexperienced newcomers to established professionals. Typically, participants took on 1500 to 3000 words of translation per project – roughly a day’s work, often with long gaps between projects. All translations were revised by a second professional. Unlike the typical setting of a translation agency, the translator and the reviser were not anonymous to each other, and communication between them and with other team members was strongly encouraged. To ensure consistency of key terms, a glossary was developed right from the first project. It was first published in 2009, then updated in 2017 (Fowler et al. 2017), with all contributors credited by name.
Some organizations emphasize the volume translated by their volunteers. For example, volunteers for TWB have the number of translated words displayed on their ProZ.com profile page (ProZ 2011), and TWB itself welcomes website visitors with a counter that shows in real time the ‘number of words donated’ (see Figure 3 above). This is very much in line with the “alternative value systems” created by social platforms such as Twitter and Facebook – consisting of likes, views, etc. – which have become “a kind of new currency” (Rushkoff 2016, p. 31). By contrast, translators in the SI network spend a considerable amount of time communicating with others, whereas in a more ‘streamlined’ setting they could presumably translate a greater volume. High productivity has never been a priority for them; in fact, productivity-maximizing strategies may easily reduce volunteers’ interest. The SI team, moreover, chose not to prioritize productivity because emphasizing the sheer number of words
5 Piróth (the first author of this article) was involved in this initiative.
contributes to the commoditization of translation. Instead, translators’ names feature in printed brochures, including the credits section in SI’s annual reports. This gives translators recognition, emphasizes the importance of translation to readers, and gives SI the assurance that participants will do their best, since their own reputation is at stake.
While many translators find that helping a humanitarian organization is rewarding in itself, the benefits of collaboration should not be underestimated. Shared projects can be the basis of future partnerships among translators. Experienced freelance professionals frequently emphasize the importance of having a trusted business partner, as it alleviates isolation, makes regular work less stressful, helps ensure a reliable backup for holidays, and may open new revenue streams by allowing modest scaling, i.e., handling projects that are too large for one person. The SI network thus serves to create relational capital for freelance translators. Shared pro bono projects for humanitarian NGOs may facilitate finding good network partners, especially because the environment is less competitive than typical translator-reviser-translation agency settings, where financial interest may turn collaborators into competitors.
The Limits of Unpaid Work
It is often considered bad taste to raise critical questions about charity and volunteering, as we have already noted, but the potential contribution of pro bono work to precariousness in the labour market is a pressing issue. NGOs in the international development field carefully consider external factors so that their actions do not harm the physical or economic environment in which they operate. This should include the community of translators who support their services. Indeed, the International Federation of Translators’ position paper on internships (FIT, 2016) stresses that “If all other members of the staff of the non-profit organisation do their work on an unpaid basis, then it is fair enough that the translator/interpreter intern is not paid either. But if other staff members are paid for their work, then there is no reason not to remunerate the translator/interpreter intern”. Naturally, this argument does not apply to interns alone.
During the Ebola crisis, TWB volunteers translated 81,000 words across all languages (Words of Relief, 2015). This is valuated at roughly
$16,000 using TWB’s usual conversion factor (USD 0.20/word), bearing in mind that the detailed program report (HIF-TWB, 2015) reveals high organizational and Machine Translation/MT-training costs, whereas costs related to translation itself are well under the USD 0.20 reference value. It is worth noting that the hypothetical per-word rate of USD 0.20/word, usually cited to tax authorities (TWB 1990, 2015, TWB 990, 2016), the general public (Kelly, 2011) and potential sponsors, is in fact several times higher than what bulk-market translation companies – which have long been represented on TWB’s board of directors and advisory board – actually pay their freelance translators, raising questions about TWB’s motives for inflating the hypothetical value.
Nonetheless, a $16,000 budget would have comfortably allowed TWB to pay its translators handsomely during the Ebola crisis. An even smaller amount would have sufficed to pay those whom TWB engaged in 2015 to “minimize the devastating effects of the Nepal earthquake” by translating, among others, “over 500 terms into Nepali, Newari and Hindi for search and rescue people and for people monitoring messages coming from the affected populations” on a volunteer basis (TWB, 2015). Importantly, TWB’s approach here is clearly at odds with the practices of humanitarian NGOs, which typically collect funds in Western countries and employ paid national staff for their aid programmes, thus contributing to the revitalization of the local economy. This is what donors expect after an earthquake that destroyed about half the country’s annual GDP.
SI attempts to avoid this ethical black hole by pursuing a different model. In 2009, it set aside a sum of €2,000 to pay its translators, consulting with them about how this amount might be shared among them. The idea of rewarding past projects was quickly discarded, as payment would have been far below professional levels and would have established an inappropriate baseline. After much discussion, a seemingly inequitable solution was agreed: to pay some participants but not others. Looking at the question from the angle of peer-to-peer solidarity led to a joint decision to reserve the limited funds available to pay colleagues without a stable income: qualified early-career translators for whom this could be the first career step.
They would probably have more time to devote to SI, so they could be expected to contribute much more than those participating pro bono. The idea of a paid remote internship was thus born. Over a period of three months, ‘interns’ – who could be located anywhere, and work from their home office – would devote ten to fifteen hours per week to SI and build their freelance career in parallel. They would be paid €1,000 each in total– a modest sum on a professional level but a decent one for the equivalent of a one-month full-time internship at an NGO in France. The rest of the team would continue to help pro bono occasionally, as their schedule would allow. Over the years, previous interns would take on more and more of the pro bono revision and mentoring work; their commitment would thus extend beyond the three-month internship period to ensure a self-perpetuating setup. These ideas were put forward in a group discussion, since a decision that would affect the overall collaboration so profoundly had to be decided collectively. The team’s unanimous approval opened the way to the annual SI internship programme. To make the internship even more useful, a series of ten two-hour webinars was created for the interns and the rest of the team on various aspects of the profession such as translation tools, business issues and revision and quality assessment protocols.
Universities often require graduating students to complete internships. Unpaid internships have become standard across the board, facilitating the recruitment process at marginal costs for companies and public institutions. Even UNICEF runs unpaid translation internship programmes – although, given the size and status of the organization, one would expect them to offer paid traineeships, as does, for example, the European Parliament’s Terminology Coordination Unit. Although SI’s paid internship programme is not a sustainable long-term career option, it is a step in that direction. If other organizations followed suit and NGOs working in the humanitarian and international development fields set similar internship conditions, for- profits would be under more pressure to improve their offers. With these considerations in mind, SI’s internship programme has been advertised at universities in the UK, US and Canada since 2016. In addition to their suitability for the task, SI’s guiding principle of solidarity rather than charity meant that candidates were also assessed in terms of how they would benefit from the internship programme.
Machine Translation and the Money Trail: Who Benefits from Volunteer Work?
As O’Donnell (2016) puts it, “[m]ake no mistake, there is big money in the international volunteering industry”, and hence “*i+t is the volunteer’s responsibility to learn about the ethical quandaries, issues, and attitudes within this industry” (bold in original). Volunteering Grassroots, the site founded by O’Donnell in 2011 “as a way to decommodify the volunteerism industry”, proposes a number of criteria for “assessing an organisation one is considering volunteering with ethically”. Prominent among these criteriais “money trail” – that is, establishing who ultimately benefits from the volunteering work.
TWB’s use of machine translation dates back to the 2010 Haiti earthquake, when Carnegie Mellon University researchers released their data on Haitian Creole (CMU, 2010). It was around this time that TWB relocated from France to the US and corporate heavyweights joined its boards en masse, expanding its management structure. President Obama’s Strategy for American Innovation 2009 had included “automatic, highly accurate and real-time translation between the major languages of the world – greatly lowering the barriers to international commerce and communication” just a couple of months earlier (Obama, 2011). In addition to international commerce and communication, intelligence (military, police, and business) also remains a high priority field of application for this technology.
In 2014, TWB received a $250,000 ‘Technology for Good’ research grant from Microsoft (TWB, 2014) to fund a crowdsourcing application to help communicate with aid workers in Swahili and Somali when disasters strike, and to build a corps of vetted translators and interpreters, plus machine translation capacity, in under-resourced world languages. In The Shock Doctrine – The Rise of Disaster Capitalism, Naomi Klein (2007) investigates how crisis situations are exploited to push through controversial policies while citizens are too distracted by disasters or upheavals to mount effective resistance. Organizations like Movement Generation thus emphasize the importance of setting up a critical framework for ‘Just Recovery’ (Movement Generation 2017), to amplify collective efforts in the face of disaster situations and make sure that they are not hijacked. Such a framework does not seem to have been considered by TWB, which used the Microsoft grant funds to verify the efficiency of the technology supplied by Microsoft itself – a company that was also represented on TWB’s advisory board. Further, the entire initiative was undertaken in response to a recommendation by the market-research firm Common Sense Advisory in its report on The Need for Translation in Africa (Kelly et al. 2012). Common Sense Advisory co-founder, Renato Beninatto, was also curiously on TWB’s advisory board at the time. In addition to the crowdsourcing application it produced, the grant also helped to set up free-of-charge ‘linguistic assets’ (human and machine), but did not, however, pay those who provided linguistic services. Acknowledging TWB’s help, Microsoft launched its Swahili translation tool, integrated into Microsoft’s various commercial products, in 2015 (Microsoft, 2015).
Similarly, speaking about the work of Translators Without Borders in Haiti after Hurricane Matthew in 2015, TWB’s chair highlighted the
organization’s issuing of cholera prevention messages and post-hurricane warnings in Haitian Creole with the help of forty volunteer translators – in close collaboration with Microsoft, with a view to improving Microsoft’s machine translation engine for Haitian Creole. The participation of TWB volunteers in building machine translation capacity for Microsoft is thus a specific example of the changing landscape of ‘charity’: here, a linguistic asset created collectively by volunteers in a humanitarian context was transferred to a for-profit project partner and turned into saleable intellectual property. This is in stark contrast with the practices of MSF, which “distances itself from the pollution of political capital and economic capital … [and] refuses to be driven by donors’ agendas in its choice of projects” (Krausse 2014,p .122).
Close examination of TWB’s structure confirms that the example of machine translation capacity-building for Microsoft by volunteers is not a one-off slip or oversight but an intrinsic feature of the way the organization operates. TWB relies on thousands of freelance translators who enthusiastically contribute on an unpaid basis, whereas TWB’s leadership has long been composed primarily of major industry players, many of whom own or operate commercial concerns that have a strong and undisguised interest in exploiting machine translation and unpaid crowdsourcing. Over the years, major users of machine translation and crowdsourcing (Facebook, Google, Microsoft, Symantec, Adobe, Oracle, MacroMedia, dotSub, Paypal, etc.), some of the largest translation companies (such as Lionbridge, Moravia, Textminded and Elanex), as well as agenda-setters of the bulk translation market (such as ProZ.com, TAUS, Common Sense Advisory, Localization World and Multilingual Magazine) have all been represented on TWB’s board of directors or advisory board, making TWB look like the philanthropic arm of a massive business consortium.
The invocation of charity and humanitarianism makes objective commentary and critique a minefield, as already noted. If Reporters without Borders had Arianna Huffington, Michael Bloomberg and Rupert Murdoch as board members, questioning their representation would only be natural. But when perceptions of conflicts of interest within TWB were raised on Kevin Lossner’s Translation Tribulations blog in October 2014, various people expressed their “*sadness and shock+ by the unjustified, small-minded and (what appears to be appallingly poorly informed) attack on an organization that has a decades-long history of providing urgently needed charity for people in dire health crises” (comment on Lossner, 2014)
– ignoring the actual issue of conflict of interest. No further comments
were added by the same critics of Lossner’s blog when specific details of the ACCEPT project were published (Piróth, 2014).
The avowed aim of ACCEPT (Automated Community Content Editing PorTal) was to enable “machine translation for the emerging community content paradigm, allowing citizens across the EU better access to communities in both commercial and non-profit environments” (ACCEPT, 2012). The project received an EU grant of 1.8 million euros, allowing the participating for-profits – including the IT giant Symantec as well as two for-profit companies, Acrolinx and Lexcelera/Eurotexte, run by board members of TWB66 – to lower their R&D costs for a disruptive technology that yields them high profit margins. In the ‘Exploitation Plan’ (no irony intended) of the ACCEPT project, Lexcelera committed itself to “scaling up the operations of Translation [sic] Without Borders from millions of words per year to tens or even hundreds of millions of words” (ACCEPT 2013). Piróth (2014) concludes the following:
Using *TWB’s+ unpaid participants in a project with an admitted commercial motive, funded by and for the EU, appears – at very least – curious. From a distance, one might ask whether TWB’s name and fame (derived from the idealistic and unremunerated contributions of donor translators focused on developing nations) has helped profit-making concerns – Acrolinx, Lexcelera, Symantec – obtain public monies for developing valuable digital media translation solutions. The ACCEPT project may yield results that justify its public funding, but they will be specifically for EU (First World) nations. TWB and other non-profits would doubtless receive some benefits, but the outcomes and assets would be ripe for use in prime commercial settings far removed from developing nations and the motivations of most volunteers.
A couple of days after the questions on conflicts of interests were raised, Lori Thicke (who founded Translators Without Borders as part of her Eurotexte translation company) stepped down, leaving TWB’s chair, curiously, to Andrew Bredenkamp, the CEO of Acrolinx.
The ACCEPT project, with massively funded digital media companies using volunteer contributors, reflects the general trend noted by Lanier (2013:257), where “network-oriented companies routinely raise huge amounts of money based precisely on placing a value on what ordinary people do online” while repositioning the same people “out of the loop of their own commercial value”. The massive scaling promised in the exploitation plan was probably a key to success; as Rushkoff (2016,p.5) notes, “[g]rowth is the single, uncontested, core command of the digital economy” and “the logic driving the low-wage gig economy” (ibid.:4). Consequently, he argues, platforms are optimized “not for people or even value but for growth” (ibid.:6).
TWB vs Humanitarian Organisations: Different Paths of Evolution
In her presentation at the 2012 TAUS European Summit, Lori Thicke emphasized the importance of “disintermediation”, of “putting the crowd in direct touch with the NGO and then getting out of the process”. Just as in the ACCEPT project, she recommended the same approach in non-profit and for-profit settings: “This is the same kind of infrastructure that I believe could be used to support other translations where there is no traditional budget, like customer support” (Thicke, 2012). But it is disingenuous to describe the aim of replacing human intermediaries by an all-logging communication platform as disintermediation, given that the idea here is to consolidate the intermediary’s role and to enable scaling and lock-in, as the example of Uber clearly shows. This slippage of terminology is not new or accidental. Morozov (2013) shows how digital media companies, under the banner of ‘disintermediation’, have introduced a growing number of mostly invisible intermediaries, a situation that might more aptly be described as ‘hypermediation’.
The emerging platform economy (Lanier, 2013) and “disruptive technology-driven productivity gains” (Kapur et al. 2005) are widely identified as key drivers of increasing inequality but are central to the operation of corporate bodies. The translation industry, represented by TWB, now offers potential investors free crowdsourcing combined with machine translation technology – on a platform that is a potential treasure trove for HR managers. In his Brussels talk, Mr. Bredenkamp mentioned that TWB would soon start collecting contributions from ‘partner NGOs’ to sustain this platform. Today, subscription fees for NGOs start at USD 500 (TWB Kató, 2017), while translators continue to contribute pro bono. TWB is not the only large-scale collaborative volunteer translation platform adopting a technology-driven approach and a top-down management structure: The Rosetta Foundation (Translation Commons/Trommons) (The Rosetta Foundation 2017) is another well- known example. For years, there has been a significant overlap between the major stakeholders of TWB and The Rosetta Foundation, including members who served on the advisory boards of both organizations simultaneously. It thus came as no surprise when the two organizationsmerged in June 2017 (TWB-TRF, 2017). Mergers and acquisitions are standard practice in the for-profit sector but are rare among humanitarian and international development NGOs. This stage in TWB’s evolution can thus be more readily understood in terms of the corporate vision of the translation industry reflected by TWB’s management and practices than by TWB’s non-profit status or chosen position as an actor in the humanitarian field, providing support for its partner (client) NGOs. Human resource management, technology and access to future EU funding are admitted key motivations: “the merger gives Translators without Borders (TWB) access to The Rosetta Foundation’s community model and technology. It also gives TWB access to EU funding through the Irish registration” (TWB Merger FAQ, 2017).
Until the beginning of this century, humanitarian and international development organizations often relied heavily on volunteers. In the past two decades, they have overwhelmingly chosen the path of professionalization, employing qualified professionals. TWB’s evolution has been quite different. As mentioned earlier, Traducteurs sans Frontières initially worked with the paid in-house translators of Eurotexte. Through a system of skills sponsorship, French companies can obtain a tax break for providing professional services to approved cultural and humanitarian organizations. This way, the French state financially supports the professionalization of these organizations.77 Traducteurs sans Frontières did not make use of this benefit, and its later transformation as the current TWB adopted a large-scale crowdsourcing approach. It now imposed professional credentials as entry criteria, but dropped payment to zero – outdoing even Amazon’s notoriously poor-paying Mechanical Turk. Demonetization usually goes hand in hand with deprofessionalization, making it particularly noteworthy that TWB managed to set up a large- scale demonetized service using professionals. This development is not in the interest of the thousands of translators who constitute the large base of the TWB pyramid, especially when professional practices are quickly being eroded by TWB’s policy of ignoring the ‘four-eyes principle’ recommendation of industry standards and skipping revision on grounds of urgency: “Since there is no time for reviewing and no room for errors in the handling of emergencies, Translators Without Borders recruits only experienced and solid professionals able to do a good job each time” (ProZ Blog, 2011).
TWB’s activity in Kenya merits a separate detailed account. Shortly after setting up a Healthcare Translation Centre in Nairobi, where hundreds of translators have been trained to date, TWB launched a ‘Fund a translator’ program on ProZ.com, targeting professional translators as potential donors. Upon inquiry (ProZ 2012), it was clarified that the program was not meant to fund translators but their training. Program director Rebecca Petras admitted that the name of the program “could be deceiving”. As TWB’s form 990 declaration (TWB 990, 2015, TWB 990, 2016) states, moreover, TWB provides financial support to TWB Kenya as “an independently registered non-profit”. Thus US-registered TWB could conveniently claim that it “did not invest in, contribute assets to or participate in a joint venture or similar arrangement with a taxable entity”
– which means it is none of IRS’s business whether Microsoft obtained any Swahili language assets in a joint venture (or similar agreement) with TWB Kenya.
TWB actively participated in monitoring elections in Kenya in 2013 and 2017, through translating communication on social media. A rapid response team of TWB provided translations into English “as quickly and accurately as possible”. A paid consultant was hired to monitor the translation process (NGOjobs, 2017), “to help determine the effectiveness of *the+ approach”. This case is not unique: TWB has created some paid positions – including a paid 3-month ‘crisis response intern’ position with a monthly stipend of USD 400 (TWB, 2018), in which the intern’s role is to support “the team on a daily basis, with a focus on managing TWB’s engagement with our community of volunteer translators during emergency responses”. While this position is certainly an interesting opportunity for someone starting out in community management, the ‘volunteer translator – paid manager’ model is highly problematic. In its FAQ section, TWB argues that NGOs often do not have a budget for translation because their core activities need to be prioritized. But how does the same argument apply to TWB itself? TWB found the necessary budget to cover substantial organizational, technical and other costs in the Word of Relief project (HIF-TWB, 2015), to monitor and assess the translators in the Kenya elections, to manage the community of translators during emergency responses, and to ensure that the participating for-profits in the ACCEPT project were handsomely paid – while those who undertook the core task, translation, were systematically asked to work on a volunteer basis. This is not an unfortunate lack of budget for translation: it is a policy decision from the top of TWB’s pyramid, which should not come as a surprise given the undisguised interest some companies represented on TWB’s board have in exploiting machine translation andunpaid crowdsourcing. It is hard to imagine that TWB would adopt the same policy if its board were composed of translators who used to “work at the coalface”, as is the case in Doctors without Borders or in SI.
TWB’s translators, through their laudable volunteer work, currently continue to serve TWB’s ‘partner’ NGOs (now more accurately called ‘clients’), which are now required to financially participate in the maintenance of TWB’s platform. In this setup, the top of TWB’s pyramid continues to benefit from excellent exposure opportunities and exceptional disruptive technology-driven productivity gains, furnished by the wide base of translators working free of charge, making it a textbook example of socialized work for privatized profit. Such a policy does not reflect the priorities of TWB’s partner NGOs (as their own policies are diametrically opposite) or TWB’s in-kind donors: the thousands of volunteer translators who continue to support the organization.
The fundamental guiding idea at SI has been that those who perform skilled work for humanitarian organizations must not risk demonetization and deprofessionalization by doing so. Rather, they should be able to make a living and grow professionally – and ultimately proceed to organizational or governance roles, if they so wish.
This is not the pathway currently in evidence with the mass crowdsourcing of translation services for humanitarian ends. Translators who consider participating in such projects, achievements notwithstanding, should be aware of how the growing focus on socialized work for privatized profit can impact them and their profession. They will be right to demand the same practices that are standard for any respected humanitarian NGO: increased accountability towards in-kind donors, with exact accounts of where volunteer translations go, and rigorous assurances that the donated or generated assets are allocated as the collaborators and public would rightfully expect. There should be conflict-of-interest policies– as again is standard for humanitarian NGOs – which should be rigorously applied to the composition of the board. After all, board members of Doctors without Borders do not come from big pharmaceutical companies but are former fieldworkers, for very good reasons.
Meanwhile, the SI experience shows that it is possible to build collaborative communities of translators capable of working directly with NGOs, bypassing the mass platforms altogether in order not just to do good for the intended recipients, but also to enhance the training, professionalization and stature of translators themselves within an overall
framework of solidarity. Ultimately, as McDonough Dolmaya argues, given the ethical questions posed by crowdsourcing and volunteerism, “including corporate reliance on free labour and the potential devaluation of translation work by the general public”, we must now address the question of how “the strengths of crowdsourcing could be leveraged to make information more widely accessible while also ensuring that users who participate are doing so as part of community-driven initiative rather than a corporate-run activity” (2018,p .354). What is needed, in other words, is a model that combines the benevolence of charitable work with the reflexivity and transformative potential of solidary action, for the benefit of both givers and recipients.