Instituting socio-technical education futures: encounters with/through technical democracy, data justice, and imaginaries

A (re)composition of encountered problems and possibilities

Artificial intelligence, automation, algorithms, and datafication are increasingly instituted across educational systems and decision-making, including generative AI, student monitoring, exam grading, data analytics. The public release of ChatGPT raised stark issues of corporate infrastructuring, the politics and scientization of automation, conflicts over ethics, and the possibility of how things could be ‘designed otherwise’ (Williamson, Macgilchrist, and Potter 2023). The global reach of exam surveillance systems during the COVID-19 pandemic provoked questions about commercial provider roles, the hidden labour of automated systems, alongside the vulnerabilities of remote studying (Selwyn et al. 2021). A grades standardisation algorithm to calculate proxy grades for cancelled examinations sparked a range of unintended consequences, protests, and public trust issues (Kelly 2021). School openings and closings throughout the crisis increasingly relied upon the politicised release and reception of cross-sectoral data analytics, while attendant COVID-19 surveillance systems escalated the powers of both government and non-government entities (Mahase 2021; Oster 2021). These socio-technical controversies highlight the multiscalar power relations shaping present and future education possibilities (Macgilchrist, Potter, and Williamson 2021). A range of conceptual and methodological constraints are provoked by these shifting power relations between diverse constituents: students, educators, communities, technology corporations, professional bodies, cross-sectoral policymakers, plus all levels of government. Methods can enact worlds, Lury (2020) argues, and attention to (re)composing how knowledge continues with, and transforms through, societal institutions is vitally needed.

There are multiple issues in education that arise from socio-technical controversies related to the use of algorithms. The first issue comes from what we know when socio-technical systems are used elsewhere. Algorithmic systems are used in high stakes areas such as policing and hiring. Problems in this use include over policing of people of colour using facial recognition technologies, and reinforcement of gender bias in automated selection in hiring practices (Noble 2018). In these areas, it has been shown that complex socio-technical systems are poorly understood by both users and those affected by the systems (Algorithm Watch 2019). Similarly, there are concerns about the capacity of educators, administrators and policy makers to understand the limitations and presuppositions of using complex socio-technical algorithmic systems in decision-making (European Commission 2021; Southgate et al. 2020). The second issue is that with the increased use of algorithmic systems, along with the identification of risks (European Commission 2021), has come awareness that ethical issues like fairness and bias are seriously underdeveloped in education (Sahlgren 2021). Additionally, with a wide range of expertise involved in developing and applying socio-technical systems, there is growing contestation over what counts as relevant professional knowledge (Pasquale 2019). There are significant disparities relating to the expertise required to understand and use algorithmic systems. For example, education experts tend not to have the technical expertise needed to evaluate and critique existing algorithmic systems (Gran, Booth, and Bucher 2021). To address this disparity in expertise, fields like human computer interaction have undertaken extensive work on making socio-technical systems more visible and accessible to non-experts (Kay, Zapata-Rivera, and Conati 2020). However, there is limited work in education on broadening expertise about algorithmic systems. As such, there is an acute need to help stakeholders, especially educators and administrators, understand both the potentials and problems with algorithmic systems in education. Amid such ‘problem spaces’ (Lury 2020) are interstices of methodological potential in relation to varying actions, publics, and contexts.

In light of these always in-between conditions, this special issue aims to inspire conceptual and methodological innovation focused on instituting socio-technical educational futures across a range of learning contexts. As automated, algorithmic, and datafied trends permeate the education sector, there is a vital need to assemble new toolboxes of methods and theories for innovative critical research “to examine educational technology as they continue to mutate, evolve, extend to new settings and expand in their (un)intended uses” (Castañeda and Williamson 2021, 11). To inform this need, we were interested in papers for this special issue that could be broadly related to exploring the potential of ‘technical democracy’ (Callon, Lascoumes, and Barthe 2009), ‘data justice’ (Dencik et al. 2019) and ‘post-automation’ (Smith and Fressoli 2021). Contributors were invited to test the conceptual and empirical horizons of collective learning, action and capabilities, such as: ‘hybrid forums’ where heterogeneous groups “can come together to discuss technical options involving the collective” (Callon, Lascoumes, and Barthe 2009, 18); the interplay of data and social justice (Dencik et al. 2019); and, post-automation, a commitment to appropriate automation technologies for “more plural relations rooted in human creativity, conviviality, and care” (Smith and Fressoli 2021, 1–2).

The range of work submitted surprised us. Not the number of conceptual developments, we had expected this would be the case. But that there were many papers that outlined attempts to help people understand and experiment with socio-technical systems in education. As such, this special issue seeks to offer Learning, Media, and Technology readers a novel set of vocabulary, resources, and ideas to spark encounters within their own education contexts characterised by democracy, justice, and sustainability. As readers will note, this issue is not about synthesis, and contriving ontological and epistemological links, but about representing the diverse ideas that are brought to bear on socio-technical controversies in education. The thirteen papers in this issue bring a rich array of conceptual tools, empirical insights, and unique approaches toward (re)composing the problems and possibilities encountered with socio-technical systems in education. We acknowledge that these papers represent scholars from the Minority World, yet we hope this special issue offers a way to connect a more diverse range of researchers and regions interested in this line of transdisciplinary inquiry. In particular, how critical and situated methods can potentially inspire new possibilities for instituting socio-technical education futures in democratic, just, and sustainable ways. Next, we provide a brief overview of these papers.

Conceptualising the limits and possibilities of justice, ethics, and discourses

The roll out of automated technologies into myriad aspects of social life has highlighted the ways algorithmic bias and automated feedback loops can reinforce existing inequalities in the areas of application. In this special issue, the focus on justice from outside of education is brought to bear on the specificity of education as a domain of policy and practice, so as to explore the fruitfulness and limits of concepts of justice, ethics, and discourses.

Against a backdrop of the recognition of algorithmic bias in education, Carlo Perrotta outlines an approach to engaging with bias through both technical and conceptual avenues. Perrotta shows how a focus on dealing with bias and fairness has become a key part of areas of computer science and statistics, including the rise of the fair machine learning field, and has been a focus of fields such as critical data studies. However, Perrotta suggests that what has been lacking in both the data and social sciences has been a focus on the key question of ‘what is a just outcome?’. Perrotta’s paper works from this gap to propose an approach to dealing with algorithmic bias using Nancy Fraser’s critiques and responses to distributive and recognitive justice. Perrotta suggests that Fraser provides a deontological approach to ethics, that allows for a pragmatic approach to participatory justice in education that can deal with the growing problems of algorithmic bias in education.

The rising interest in ethics and technology in education is carried through into Jeremy Knox’s paper. Knox suggests that ethics are becoming predominant in the proliferation of new frameworks to govern the use of artificial intelligence in education. However, for Knox this means there is a hollowing out of ethical meaning that provides legitimacy for the use of technology, rather than any type of substantive guidance for not only how but whether to use emerging technologies like AI in education. Knox suggests that we need to locate the use of AI in education both within a discussion of the philosophy of ethics and in relation to the existing concerns about justice from within education. This would allow for questions about how technology can amplify or reshape existing education practices.

Janine Arantes and Mark Vicars take an approach to data justice by examining how inclusion and exclusion of data categories have significant implications for LGBTQI+ people and attendant discourses. They examine how big data and automated technologies exclude LGBTQI+ people, undertaking a genealogical approach to a series of socio-technical imaginaries that highlight the ways LGBTQI+ people as ‘missing in action’ in data. As a counter-narrative to this absence Arantes and Vicars propose that we can consider a queering of big data and AI, that is to consider ‘queer automation’.

Democratising technologies via pluralistic approaches with diverse stakeholders

This next group of papers highlights ways to potentially democratise technologies via collective learning and pluralistic possibilities with diverse stakeholders. In particular, they articulate unique modes of cooperation which engage with the uncertainty, controversy, conflict, complexity, and possibilities of emerging technologies. Each of these papers offers a unique way to engage stakeholders with diverse expertise - with the common purpose of addressing the limits and possibilities of socio-technical systems in education via pluralistic approaches.

To expand public pedagogy with emerging technologies across society, Teresa Swist, Justine Humphry and Kalervo Gulson introduce a ‘pedagogic encounters toolkit’ to study algorithmic system controversies. This multifaceted approach includes learning with controversies, testing diverse methods, and making democratic designs. The toolkit is tested in the context of school-based and urban-based cases, which highlights the potential of applying and adapting the toolkit across diverse sectors.

Within the context of socio-technical controversies in education Greg Thompson, Kalervo Gulson, Teresa Swist and Kevin Witzenberger explore the potential of 'hybrid forums' to broaden expertise and participation about the impacts of automated decision-making (ADM) systems in education. Proposed as ways of decision-making that do not require consensus, these forums can create moments of democratisation via processes of learning with shared uncertainty, material politics, and collective experimentation.

As a critical response to what they view as the limits of hybrid forums, Jessica Holloway, Steven Lewis and Sarah Langman argue for the need to embrace democratic dissensus via a process of ‘technical agonism’. Informed by Habernasian theory and a hypothetical case study focused on the Texas Teacher Evaluation and Support System, they propose the need to centre dissent and scepticism in the pursuit of democratic ideals in the context of the datafication of education.

To explore the possibilities and limits of the growing field of explicable artificial intelligence (XAI) in education, Robert Farrow combines a socio-technical perspective with the application of artificial intelligence (AI) in education to address structural issues of transparency. Farrow proposes that educators and learners should be engaged in the process of meaningfully understanding and consenting to AI interventions through an ongoing process of trust-building and interpretability.

The powerful role of social research in making educational futures is the focus of a paper by Kenneth Horvath, Mario Steinberg and Andrea Isabel Frei. Informed by critical data studies in education and French pragmatic sociology, the authors aim to bridge inquiry and critique via their analytical anchor point of ‘plural (school) worlds’. They examine the historical entanglements of social and political orders and emphasise the need to expand discursive spaces and shift problematizations amid the context of existing school worlds.

Investigating alternatives to data and commercially-led visions of education futures

This final group of papers illustrate ways to investigate alternatives to data and commercially-led visions of socio-technical education futures. Attention to the temporality of socio-technical systems in education can help open up the scope of research approaches. As technologies become increasingly networked and proprietary, expanded lines of inquiry are necessary to surface these dynamic tensions associated with datafied platforms, marketplaces, and strategies in education.

To surface the often hidden circulations of data associated with the Australian primary school context, Tiffany Apps, Karley Beckman and Sarah Howard applied a walkthrough method to a study of two digital reading platforms. This method enabled the researchers to examine the visibility of users, (mis)representations of learning, plus the (re)configuration of users, classroom teaching practices, and digital labour. The researchers suggest that this approach offers a way to generate heuristics and representations that enable multiple stakeholders to understand digital economies and engage with questions of data justice.

Amid the conditions of an increasingly commercialised curricular marketplace in the United States, Michael Brown, Noreen Naseem Rodríguez and Amy Updegraff explore the potential of alternative models of online curriculum and instructional resources. The researchers conducted a qualitative critical content analysis of how teacher candidates in an elementary social studies methods course encountered and experienced the Teachers Pay Teachers online marketplace. They propose a ‘curricular platform cooperative’ conceptual framework that would provide opportunities for curating social justice instructional content with embedded affordances for user contributions and feedback.

Also deploying a data justice lens, Molly Stewart, Elizabeth Pier, Dan Ralyea and Andrew Rice examine interoperability initiatives in several U.S. states to explore the role of data standards and APIs in the K-12 education sector. Their research exposes the power dynamics among vendors, states, and local agencies and potential data injustices associated with data access and visibility, plus the representation and categorisation of students. With current interoperability initiatives currently driven by policymakers and technology vendors, this article highlights valuable avenues for wider public learning and engagement informed by justice-led approaches.

In working towards sustainable and reflective digital school development, Nina Brandau and Samira Alirezabeigi deploy critical design (CD) and participatory design (PD) in the context of daily practices at two elementary schools. Guided by these approaches, the analysis and interpretation of ethnographically collected material identified how CD and PD challenged, or supported, different stakeholders’ practices and understandings of digital school development. While recognising the limits of time and educational administration constraints, Brandau and Alirezabeigi propose the potential of re-thinking CD and PD to grapple with tensions between immediate and long-term expectations, digital and analogue practices, plus varying modes of knowledge acquisition across daily school contexts.

In a close-up examination of high school students’ data modeling practices, Shiyan Jiang, Hengtao Tang, Cansu Tatar, Carolyn Rosé, and Jie Chao illustrate the potential of data modeling practices to support democratic discussions and informed evaluations of AI technologies. Drawing upon technical democracy and situated learning approaches, they identify practices associated with students’ data modeling processes, plus learning opportunities that could support critical evaluations of automated decision-making. Based on their findings, the researchers highlight implications for understanding data justice, the potential of designing accessible data modeling experiences, plus the role of data modelers.

Toward a collective mode of instituting

In closing this editorial, we return to the ‘instituting’ aspect of this special issue’s title. In his book The Imaginary Institution of Society, Castoriadis (1975) argues that possibilities for alterity and transformation always emerge from “a new mode of instituting and a new relation of society of individuals to the institution” (373). As the global controversy of ChatGPT highlights, the dominance of Big Tech is expanding due to a range of data, computer power, and geopolitical advantages which requires more expansive obligations, policies, strategies, and changes across society (Kak and West 2023). To counter such dominant narratives, designs, and infrastructures, there is an urgent need to make spaces for generating future alternatives and transitions across education (Macgilchrist et al. 2023). Embedding transitions with/through conditions of corporatised, infrastructural dominance across society demands a new mode of instituting socio-technical education futures.

The papers in this special issue (re)compose a range of encounters with/through technical democracy, data justice, and imaginaries. In particular, how to: conceptualise the limits and possibilities of justice, ethics, and discourses; democratise technologies via collective learning and pluralistic approaches with diverse stakeholders; and, investigate alternatives to data and commercially-led visions of education futures. As Castoriadis (1975) notes, institutions are never static, but continually dynamic in terms of how they are created, maintained, altered, or destroyed: “it is the union and tension of instituting society and of instituted society, of history made and of history in the making” (108). How we collectively grapple with these instituting unions and tensions signal both political and compositional turning points for transdisciplinary research, policy, and practice.