Data broadly, and particularly ‘big data’, is increasingly leveraged to develop a wealth of digital projects and products across academia, government and the commercial sector. While often celebrated for being ‘frictionless’ and enabling efficiency through automation, these processes of computation and digitisation implicate human labour at every stage of the process, from the production and maintenance of physical infrastructure to data extraction to coding and technical design. This labour is often either outsourced or crowdsourced. Crowdsourcing, or the mobilisation of a large community to ‘undertake a specific task, create content, or gather ideas’, has been integral to the expanding field of data science (Terras 2016). On one hand, it has been used to develop important digital tools and collective repositories of information, for instance Wikipedia and various digital archives used in the heritage sector. On the other hand, crowdsourcing labour is ripe for exploitation and erasure, particularly along the lines of class, race and gender (D’Ignazio & Klein 2020).
In order to ensure the human effort involved in ‘data work’ isn’t rendered invisible or undervalued, the authors of Data Feminism (2020) have identified “show your work” as a feminist intervention – which entails taking a data product and unpacking the material conditions of its existence. Similarly, DH scholar Miriam Posner has argued that we should look ‘under the hood’ to ‘scrutinise data, rip it apart, rebuild it, reimagine it’ (Flanders 2018). One example of this is Kate Crafword and Vladan Joler’s 2018 artwork Anatomy of an AI System, which deconstructs an Amazon Echo voice assistant through a complex commodity chain of labour, data and material resources. In addition to the planetary costs of mineral extraction and the labour needed to perform this, Amazon’s labour force comprises layers of workers, often in developing countries, who do a variety of paid and unpaid labour. This includes Amazon’s controversial crowdsourcing platform Mechanical Turk (MTurk), which is leveraged to train datasets and do network maintenance.
From a user perspective, labour involved in processes of digitisation has the potential to automate tasks which might otherwise exclude marginalised groups or reinforce gendered and raced divisions of labour. For example, one reading this week commended Trove, the National Library of Australia’s digital archive, for providing new avenues for historical archival research for people with disabilities, thereby promoting inclusivity within the discipline (Pikó & Brett 2021). However, these claims of automation as liberation from alienating work should be interrogated. For example, gendered voice assistants like an Amazon Echo come with the promise of offloading some of the monotony of housework. Many scholars have challenged this, arguing instead that this automation diminishes the household affective labour traditionally performed by women, particularly women of colour, by making voice assistants replaceable. It also allows this labour to be harnessed without actually affording labour rights or financial compensation to real workers (Schiller & McMahon 2019; Sweeney 2021). Crawford and Joler remind us that as consumers, we do the work of providing Amazon with valuable training data of verbal commands, queries and responses. Some have even argued that the feminine personas of voice assistants are employed in service of surveillance capitalism and the accumulation of data for corporate profit (Woods 2018).
While many of the processes of creating, training and operating an Amazon Echo remain largely a black box, this case study illustrates the indispensability of human labour found throughout systems which are designed to replace it.
References
Crawford K and Joler V (2018) Anatomy of an AI System. In: Anatomy of an AI System. Available at: http://www.anatomyof.ai (accessed 8 March 2022).
D’Ignazio C and Klein LF (2020) Show Your Work. In: Data Feminism. Strong Ideas. Cambridge, MA, USA: MIT Press.
Flanders J (2018) Building Otherwise. In: Elizabeth M. L and Wernimont J (eds) Bodies of Information: Intersectional Feminism and Digital Humanities. The University of Minnesota.
Pikó L and Brett A (2021) Trove, disability, and researching history: or, digital materialism for precarious times. History Australia 18(4). Routledge: 855–858. DOI: 10.1080/14490854.2021.1993748.
Schiller A and McMahon J (2019) Alexa, Alert Me When the Revolution Comes: Gender, Affect, and Labor in the Age of Home-Based Artificial Intelligence. New Political Science 41(2). Routledge: 173–191. DOI: 10.1080/07393148.2019.1595288.
Woods HS (2018) Asking more of Siri and Alexa: feminine persona in service of surveillance capitalism. Critical Studies in Media Communication 35(4). Routledge: 334–349. DOI: 10.1080/15295036.2018.1488082.