This PhD focuses on new forms of labour relations emerging within global digital capitalism. For capital to hit the ground (Mezzadra & Neilson 2019), it needs certain tools to be able conduct operations so that extractions can take place. With regard to high-tech capital, numerical interfaces that act as mediators and measurement tools are thus of central concern. This research is empirically embedded along the political landscape of refugee camps. In this specific context, the notion of high-tech labour camp is used. There, new matrixes of computation integrate with the economy of war and produce new computational measurements of labour time while they force the constitution of a new global refugee workforce. Tensions between capital and migrant mobility crystallize in high-tech labour camps (Anesse 2006, Mezzadra 2010, Amar 2013).
Specifically, biometrics is analysed as a tool to control populations and calculate labour time in the gig economy, following the theory that a machine is also an instrument to measure labour (Pasquinelli 2019). The humanitarian rationale of urgency, something must be done, and its immunity allows UNHCR to conduct political, medical and policing tests (Jacobsen 2015). Historically, new technologies and experiments were always tested and carried out on minorities or on groups perceived as inferior. Camps thus serve as political-juridical grey areas, characterised by regimes of exception and marginalisation (Agier 2011). This research thus proposes to investigate how refugees of the global peripheries become experimental, precarious populations in the camps that serve as laboratories. In these “lab camps” statistic, algorithmic and biometric technologies are tested and choreograph their performance. This computational experimentation returns, when made stable, to the Global North in the form of governance- and labour management tools.
The analysis also covers how gig-work is both invisibilised and mediated through the technical composition of AI. There’s a stark difference between the common perception of current high-tech (AI, machine learning) as a western invention and the actual relations of production that clearly show how fundamentally important non-western labour and knowledge was and is in order to get it to work and to hit the ground. Some of the most advanced technological regimes took and currently are taking shape outside the so-called developed West (Breckenridge 2014). Capitalist expansion as extraction of data and labour and its accumulation in the global North should be unraveled as a continuation of colonial forms of exploitation. Old asymmetric distributions of power and capital are echoed in more recent forms of epistemic colonialism in AI (Pasquinelli 2020). Current discourses such as “Data 4 Development”, shaped by key figures such as United Nations, World Economic Forum and UNHCR that frame these activities as “humanitarian” and therefore self-evidently as “for the good” need to be examined critically (Mann 2017).
The PhD project covers case studies that focus on biometric enrolment and refugee gig work in Jordan, North Iraq and Uganda where refugees in camps label training data for machine learning algorithms that are then used to automate complex perceptional tasks like vision. Clearly visible, automation does not replace, but rather displaces labour onto a new refugee workforce; shifts it to the minds of a double-precarious low-cost workforce. It is here where we find the hidden human faces of automation (Irani 2016). The notion of the high-tech labour camp as laboratory could thus serve as a helpful notion to understand how frictions and struggles in today’s global arena of armed conflict, war and environmental migration give shape to the technical composition of AI.
- Anesse, Aneesh, Virtual Migration: The Programming of Globalization, Duke University Press, 2006.
- Breckenridge, Keith, The Biometric State: The Global Politics of Identification and Surveillance in South Africa, 1850 to the Present, Cambridge University Press, 2014.
- Irani, Lilly (2016). “The Hidden Faces of Automation.” XRDS: Crossroads, The ACM Magazine for Students 23, no. 2: 34–37.
- Mann, Laura (2017). “Left to Other Peoples’ Devices? A Political Economy Perspective on the Big Data Revolution in Development”. Development and Change 0(0): 1–34, International Institute of Social Studies.
- Mezzadra, Sandro, Nielson, Brett, Politics of Operation. Excavating Contemporary Capitalim, Duke University, 2019.
- Mezzadra, Sandro, The Gaze of Autonomy. Capitalism, Migration and Social Struggles, Uninomade, 2010.
- Pasquinelli, Matteo, “On the Origins of Marx’s General Intellect”. Radical Philosophy, 2.06, winter 2019.
- Pasquinelli, Matteo and Vladan Joler, “The Nooscope Manifested: Illuminating AI as Instrument of Knowledge Extractivism”, AI and Society (21 November 2020).
Further Reading on the case studies:
- Ariana Dongus, Christina zur Nedden, ZEIT ONLINE: https://www.zeit.de/digital/datenschutz/2017-12/biometrie-fluechtlinge-cpams-iris-erkennung-zwang, 15. 01. 2018.
- Ariana Dongus, Christina zur Nedden, FRANKFURTER ALLGEMEINE SONNTAGSZEITUNG, http://www.faz.net/aktuell/wirtschaft/re-coded-bildet-fluechtlinge-im-irak-zu-programmierern-aus-14993418.html, 15. 01. 2018.
- The case of the Rohingya refugees enrolment in Bangladesh is but one example of the perils UNHCR’s Biometric Identity Management System imposes on vulnerable populations. See: Zara Rahman, IRIN NEWS, https://www.irinnews.org/opinion/2017/10/23/irresponsible-data-risks-registering-rohingya, 15.12. 2017 or Elise Thomas, WIRED UK, http://www.wired.co.uk/article/united-nations-refugees-biometric-database-rohingya-myanmar-bangladesh, 15. 03. 2018.