Auditing African Content Moderators’ Working Conditions by Using the European General Data Protection Regulation (GDPR) \conffull ( \confshort ), \confdate , \confloc Conference: The 2026 ACM Conference on Fairness, Accountability, and Transparency; June 25–28, 2026; Montreal, QC, Canada The 2026 ACM Conference on Fairness, Accountability, and Transparency (FAccT ’26), June 25–28, 2026, Montreal, QC, Canada DOI: 10.1145/3805689.3812228 ISBN: 979-8-4007-2596-8/2026/06 CCS: Social and professional topics Computing / technology policy CCS: Security and privacy Social aspects of security and privacy

Mariame Tighanimine, Jessica Pidoux, Sonia Kgomo, Kauna Ibrahim Malgwi, Richard Mwaura Mathenge, Mophat Okinyi, James Oyange · 2026

Read Auditing African Content Moderators’ Working Conditions by Using the European General Data Protection Regulation (GDPR) \conffull ( \confshort ), \confdate , \confloc Conference: The 2026 ACM Conference on Fairness, Accountability, and Transparency; June 25–28, 2026; Montreal, QC, Canada The 2026 ACM Conference on Fairness, Accountability, and Transparency (FAccT ’26), June 25–28, 2026, Montreal, QC, Canada DOI: 10.1145/3805689.3812228 ISBN: 979-8-4007-2596-8/2026/06 CCS: Social and professional topics Computing / technology policy CCS: Security and privacy Social aspects of security and privacy by Mariame Tighanimine, Jessica Pidoux, Sonia Kgomo, Kauna Ibrahim Malgwi, Richard Mwaura Mathenge, Mophat Okinyi, James Oyange (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-11699.txt.

Introduction

Tighanimine and six others audit African content staff working terms using the EU GDPR. They examine staff in Kenya and Nigeria employed by outside firms. Using the GDPR's cross-border scope, they gain access to contracts and NDAs never given to the workers. The approach yields legally grounded evidence of harms facing Global South staff, whose harsh terms violate workers' rights. The work also shows data-rights law against tech trade special claims, which hides reliance on outside firms to shift out labour costs and blame.

Catalog record

CREATORMariame Tighanimine, Jessica Pidoux, Sonia Kgomo, Kauna Ibrahim Malgwi, Richard Mwaura Mathenge, Mophat Okinyi, James Oyange
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE61696
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.11699

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Auditing African Content Moderators’ Working Conditions by Using the European General Data Protection Regulation (GDPR) \conffull ( \confshort ), \confdate , \confloc Conference: The 2026 ACM Conference on Fairness, Accountability, and Transparency; June 25–28, 2026; Montreal, QC, Canada The 2026 ACM Conference on Fairness, Accountability, and Transparency (FAccT ’26), June 25–28, 2026, Montreal, QC, Canada DOI: 10.1145/3805689.3812228 ISBN: 979-8-4007-2596-8/2026/06 CCS: Social and professional topics Computing / technology policy CCS: Security and privacy Social aspects of security and privacy Mariame Tighanimine, Jessica Pidoux, Sonia Kgomo, Kauna Ibrahim Malgwi, Richard Mwaura Mathenge, Mophat Okinyi, James Oyange Abstract . In this article, we audit the working conditions of content moderators in Kenya and Nigeria employed by business process outsourcing (BPO) companies by using the European General Data Protection Regulation (GDPR). We demonstrate its extraterritorial scope for gaining access to elements such as employment contracts and NDAs that have never been provided to the workers concerned. The results of this approach provide legally grounded evidence of the structural disadvantages faced by content moderators in the Global South, whose exploitative working conditions violate workers’ rights. Our work also highlights the benefits of legislation aimed at protecting individuals’ data rights as a counterweight to the tech industry’s discourse of exceptionalism, which obscures its dependence on BPOs to externalise labour costs and accountability, whilst claiming that its products, business models, and methods of resource extraction are unprecedented and fall outside any existing legal framework. † † cc-license: by 1. Introduction The systematic concealment of industrial labour within supply chains by companies has persisted throughout history, obscuring workers from public view and enabling the precarisation of their conditions, while simultaneously giving rise to collective demands and social progress. The precarious situation of content moderators in the contemporary AI industry is not unprecedented; rather, it represents the latest manifestation of this longstanding dynamic. Yet the technology industry also exhibits distinct characteristics that warrant specific attention. In recent decades, a narrative of exceptionalism has become prominent in this sector. This socio-legal imaginary treats platform companies, their products and associated business models as disruptive and exceptional, thereby enabling them to engage in regulatory arbitrage ( 39 ; 116 ) . Critically, for workers, this implies that they do not fit any form of existing working models. However, numerous studies have shown the precarity, retrograde and exploitative nature of the forms of work in platforms like Uber ( 17 ; 50 ; 86 ; 87 ; 18 ) and more generally with actors in the gig economy over the world ( 9 ; 6 ; 76 ; 110 ) . Recent gig worker’s mobilisations ( 109 ; 121 ; 34 ) and the adoption of legislation like the EU’s Platform Workers’ Directive have proven that platform exceptionalism relates more to a well-honed narrative to justify an economy of servitude ( 74 ) . In the AI industry, the exceptionalism narrative has gained renewed vigour. The development of increasingly powerful AI models that are presented as exceptional in human history ( 16 ) —the race toward AGI ( 21 ; 40 ) represents the strongest version of exceptionalist discourse—has been accompanied by reliance on low-wage labour, particularly in countries in the Global South ( 10 ) . In recent years, the demand for data ( 44 ) —in a culture of machine learning where data work is seen by computer scientists as “silly work” ( 51 ) and working on models is perceived as real work—has intensified, considered as a form of “voracious appetite” ( 25 ) to the point where questions about compliance with key privacy principles are being raised. Consequently, this raises the question of whether technological exceptionalism serves to justify a regression in fundamental principles and rights at work as defined by the International Labour Organization (ILO). The aim of our work is to examine the tangible effects of these developments. By focusing on a collective of content moderators based in Africa who have formed a union to defend their interests as workers, denounce their working conditions, and gain recognition for their professionalism, we have co-led participatory research demonstrating that African content moderators are deprived of conditions that guarantee “decent work” ( 122 ) or more accurately, “meaningful work” ( 92 ; 93 ; 91 ) . These workers are denied knowledge of the content of their work in their employment contracts, which is an essential element of the social bases of self-respect ( 91 ) . They also lack access to a clear and explicit description of their tasks, which would enable them to decide how to perform their work and have a say in the evolution of work processes and company policies ( 13 ) . Finally, workers were required to sign particularly restrictive non-disclosure agreements (NDAs), of which they were not provided a copy. The institution of the employment contract is considered a major step forward in the historical transition away from forced labour. It made possible the freedom to participate in the labour market, which Sen ( 99 ) considers both a condition for development and a fundamental freedom. Our methodological approach to deconstructing technological exceptionalism draws on an analysis of content moderators’ working conditions, using the European General Data Protection Regulation (GDPR) as an investigative legal tool. By facilitating workers’ exercise of their data subject rights, this approach enables an audit of AI technologies from the perspective of those who perform data work: a crucial yet often overlooked component of the AI supply chain. Our work contributes to the research that has been developing in recent years to address issues of fairness, transparency, and accountability in AI systems in a concrete manner. Indeed, precarious and hidden human labor is an integral part of AI systems ( 123 ) . Thus, “any serious conversation about the ethics and justice of AI must include the workers who make these systems possible ( 61 ) .” We strive to be consistent on ethical issues, contributing both to documenting the conditions under which AI systems are produced by human workers and to advancing research that directly engages those workers. This study makes two contributions. First, it is methodological and legal: it demonstrates that workers in the AI supply chain, located in jurisdictions with weak data protection frameworks, can invoke GDPR rights against employers whose data processing activities involve transfers of personal data to Europe, in order to recover employment documents and audit their working conditions. Second, it is empirical: drawing on four cases, we provide legal evidence on working conditions derived from documents transmitted by the companies themselves, offering a rare insider view of labour practices in the AI supply chain. 2. Context and Motivation: From Coal Mining to Data Work What content moderators experience today, including invisibility, job insecurity, low pay, and lack of unionisation ( 10 ; 26 ) , is comparable to what coal miners experienced in the early days of industry. Both industries are fundamentally extractive: coal mining extracted raw material to power industrial capitalism, while the technology industry extracts data processed by workers to power AI systems. This labour is substantial: Muldoon et al. ( 68 ) note that data processing, defined as “human labour required to support machine learning algorithms through the preparation

Read the complete text