Brave new platforms: a possible platform future for highly decentralised schooling

ABSTRACT

Sweden has one of the most marketised and decentralised school systems in the world while also ranking amongst countries with the highest levels of access to technology in classrooms. Considering the increasingly central role that digital platforms play in the practices of schooling, this article speculates on what might happen during the 2020s in highly decentralised school systems like Sweden’s. Based on current trends in education, directions indicated by platformisation in other contexts and taking a critical speculative approach, it offers a discussion of what could happen to the practices of schooling and the public mission of education. This discussion is intended to raise important questions for researchers, educators and policy makers to consider as the platformisation of schooling unfolds.

KEYWORDS:

• Education
• school
• platforms
• data colonialism
• speculative approach

As early as 1912, Thorndike speculated about a future where mechanical marvels could do the work of teaching. Since then, the idea that machines could replace teachers and reorganise schooling has emerged again and again as each successive wave of ‘teaching machine’ technology has formed, crested and eventually broken against the powerful institution of schooling. However, as a new wave of teaching machines begins to form with the rapid digitalisation of the practices of schooling, advances in data harvesting and the advent of artificial intelligence (AI), we ask what might happen in a context where schooling as an institution has already been broken-up, decentralised and marketised?

Similar to other Organisation for Economic Co-operation and Development (OECD) countries, Sweden is experiencing a tremendous push for school digitalisation. However, unlike other countries, until the 2017 publication of a high level national visionary policy (Department of Education 2017) the push for digitisation over the past few decades was not based on strongly regulated strategies for school digitalisation or national regulations for data archiving or infrastructure. Rather, the push was a response to and mainly a part of a rapid decentralisation and a radical marketisation that allows for-profit schools (Lindvall and Rothstein 2006; Lundahl 2016). As a part of this, a widespread infrastructural push of one-laptop-per-child programs across municipalities and schools has taken form. These changes have yielded a situation with little state governance where the dominant technical platforms are amongst few centralising powers uniting schools as a national school system. Sold as ways to organise and administer schooling, these platforms are an increasingly dominant influence on how Swedish schooling is accomplished and may be becoming a de facto regulatory regime.

Given the unfolding situation in Sweden, the question we wish to explore here is what could happen to the practices of schooling and the public mission of education in a system so susceptible to commercial influence and platform logics? Based on the conditions of the current Swedish model, recent occurrences within the education sector and directions visible in relation to platformisation in other contexts, we will offer speculative accounts of future developments in education when both the data itself and the decision-making at national, school and classroom levels becomes owned by commercial interests and is performed by advanced, adaptive algorithms.

To be able to speculate about a future socio-technical development of this kind, we make use of a critical and mainly pessimistic approach. We are aware that we are culturally attuned to see stable and inevitable historical lines for how adaptive digital platforms will develop. However, we try to avoid taking a deterministic techno-pessimistic position and instead our method is to extrapolate and exaggerate relevant current and historical events as a form of case ‘evidence’, taking diagnosis of the present as a starting point. Our purpose is to contribute to a discussion about the potentially problematic consequences of current and readily conceivable adaptive digital platforms. For example, we relate the development of such platforms to the current conditions of commercialisation and marketisation in the Swedish public education sector, but we also bring in other perhaps less obvious or visible aspects to our speculations. For our analysis, we make use of the current state of the case and then aim to historicise around three possible near futures. We present these speculations as narratives related to a processual development consisting of three stages. As proposed by Jen Ross (2017, 227), ‘speculative approaches need to produce findings and conclusions which are generative enough to serve as invitations for further work, and indeed to offer insights for educational practice.’ Our hope is that the speculative analysis presented here provides insights for educational practice and how it relates to digital platforms. In line with Cathy O’Neil (2017), we wish to raise questions around the responsibility of algorithmic powers and how platform use is (un)regulated by foregrounding how developments and (non-)decisions in the present may have unintended consequences in the future.

Our analysis relates to the emerging field of analyses of digital platforms (e.g., Van Dijck, Poell, and de Waal 2018) that examines the ways they regulate and govern by ‘datafication’ practices. In particular, we draw on discussions relating to digital platforms that reveal how they govern public spheres like the education sector in different ways (Gillespie 2010). A key concern is that global commercial platforms incorporated into public education risk challenging education as a public good (Poell et al. 2018). Digital platforms are described as ‘online content-hosting intermediaries’ (Williamson 2017a, 62) with designs that allow interaction between a ‘platform core’, programs and user-generated content through application programming interfaces. This arrangement makes it possible for platform providers to profit on the data produced by users, a process which then re-shapes education ‘by the business model and political economy of platform capitalism’ (Williamson 2017a, 62). In this way, essential and possibly sensitive aspects of schooling and the digital lives of the individuals involved are made available for commercial exploitation. Furthermore, the algorithmic powers of digital platform data production not only make it possible to predict and profile individual user behaviour, but also to regulate and differentiate groups of users in intricate and unpredictable ways, black-boxed by the technology (Couldry and Mejias 2018; Macgilchrist 2019; Williamson 2017b). Such factors already have consequences in educational contexts with platforms themselves becoming templates for how education is arranged (Williamson 2017a). However, as the centrality of digital platforms increases, it is reasonable to assume that the near future will see an increase in the seriousness of consequences for the mission of public education and the practices of schooling. In our efforts to further the discussion of such future consequences, we begin with the case of Swedish schooling as the year 2020 begins.

Stage 1: schools and digital platform marketisation

David and his brother had always been somewhat competitive. When they both had children around the same time, they continued their usual quarrelling by comparing them. David had been trying to ignore his brother’s bragging, but at a family gathering to celebrate the new year, he heard his brother say something that caught his attention. It was something about how great his son was at writing. He was describing how a system from the search giant Poodle that his son Martin uses for school automatically corrects his writing and changes the exercises he does based on the errors he makes. ‘I didn’t even know that Poodle made stuff for schools’, thought David. His daughter Karin had been doing the same exercises on her school computer, but the software had the school company logo in the corner. Surprised, David suddenly spoke up, ‘Wait, do you guys have the same system? Is it Poodle?’

Beginning in the late 2010s, Poodle carved itself a dominant position in the school platform market in Sweden. The search giant began by offering administrative features, then moved into offering tools for classroom use that exploited its many services from video to word processing. In the years that followed, a new digitally oriented curriculum was introduced and with its dominance in the market for provision of digital services firmly entrenched, Poodle began to offer more and more instructional content. Through a push into AI in the middle of the decade, Poodle started to offer adaptive instructional technologies like writing support and test marking. At the time of its introduction, the Poodle School platform branded for each school company raised few eyebrows. Few parents even knew that their kids had started to work on a different platform in their schools. The company was American and generally considered to be a responsible corporate citizen. And, after all, it provided a rich suite of tools for teachers and students to use at a low cost. It wasn’t until much later that it would become clear that the unique conditions of the Swedish school system combined with the integration of digital platforms into virtually all aspects of schooling would lead to some unforeseen issues. Issues that relate to particular events that took place several decades earlier.

Poodle logo.

In the 1990s, Sweden embarked upon an aggressive reform of the school system that drastically decentralised control over what was taught and how resources were distributed. This decentralisation was combined with an extreme marketisation that by the 2010s left Sweden as the only country in the world that allowed state-funded for-profit schools. In this economic landscape, companies acquired schools as a way to generate profits from state funding, often very successfully. The number of organisations running schools blossomed and the combination of aggressive marketing by for-profit school companies and freedom for parents to choose which school their children should attend became a well-established practice of schooling. While this shift was initially widely accepted, concerns were raised by the end of the 2010s that the downsides to decentralisation, school choice and for-profit education such as increasing segregation were becoming prevalent. A conflict emerged between the (neo-)liberal value of freedom-of-choice and the socialist egalitarian principle that was a dominant concept in Swedish politics during much of the twentieth century. However, an increasingly populist orientation in the politics of the late 2010s made a return to a school system without the same freedom-of-choice or possibility to run for-profit schools politically unpalatable.

For a period in the mid-2020s, the vast majority of schools were privately run by an industry of over 300 different companies. With free hand to choose how instruction would be provided and pressure from investors to increase profit margins, these companies enthusiastically adopted the services and instructional content that digital platforms, predominantly Poodle School, provided. The adaptive features enabled by AI technologies that were successively introduced to these platforms were accompanied by arguments for their efficacy based on the vast data traces that were produced and directly connected to an ever-growing range of global assessment standards that were primarily managed by the OECD Programme for International Student Assessment (PISA). With the entry of several large business intelligence and analytics providers into the education market who offered real-time indicators based on PISA models and direct access to platform data, platform providers like Poodle together with the developers of plugin modules for their ecosystem were able to formulate very strong data-driven arguments for the mass introduction of their products. With strong arguments about efficacy in terms of student attainment, the adaptive features of platforms became a way to reduce the number of teachers needed and soon the digital platforms, their content, guidance and assessment became the de facto curriculum, and the technology companies the de facto ministry of education. In an otherwise fragmented situation, the dynamic platforms became the one thing that schools in Sweden had in common.

Stage 2: platform data exploitation

Erik and Anna first came into contact with DifSchool when they had to choose a school for their daughter during a one-year period working abroad. Anna had got a job as foreign correspondent covering South America for Swedish radio news. Since Erik had taken a leave of absence from his job, the couple decided to enrol their daughter in DifSchool, a cyber charter school system run by Poodle. With Erik able to home-school, the support DifSchool provided had worked well and the couple were satisfied with their experience. When the couple returned to Sweden in 2025, they found that DifSchool had acquired several of the schools in their local area and since their daughter’s data was already on the Poodle platform, she could easily continue where she had left off.

Around 2020 most of the commercially run schools in Sweden began to show declining profit margins. The market had become saturated and rationalisations to reduce costs such as increasing student to teacher ratios and decreasing teacher contact time with students could not be stretched much further. At this time, regulations regarding ownership were changed with the market opened up for international companies to own and run Swedish schools. This would allow Poodle to change its strategy and tighten its grip over the sector. Through their subsidiary company, DifSchool, they began to buy up schools and school consortia. The DifSchool company had started based on the American model of cyber charter schools in the early 2000s and by the late 2010s had achieved a global outreach (Courtney 2018). As they received the same state funding as other schools, but had reduced overhead costs with their digital technology-based education and home-schooling approach, the cyber charter school model proved very profitable. Added to this, companies could sell the model to policy makers as a ‘public good’ that could provide equitable access to the curriculum for all, including those in remote areas such as northern Sweden.

DifSchool logo.

While there were clear opportunities to make large profits through the cyber charter school model through their DifSchool subsidiary, Poodle chose a different approach in Sweden. By the early 2020s, Poodle had adopted an aggressive new tactic that actually meant ploughing more money into the schools than they generated. By hiring more qualified staff and offering attractive locations, they soon became the dominant educator in the nation. Unlike the previous owners of these schools, their interest was not in short-term profits, but in the data that the platforms used could generate. Data on all aspects of students were monitored, not only basic measures like school performance and attendance, but also other socio-emotional and biological aspects such as their emotional status, social life and genome profile. This shift coincided with the launch of a new global PISA programme, now operated by one of the largest multinational education companies. The new assessments were heavily imbued with AI technology, automating many of the performance and prediction criteria already being performed digitally on platforms and plug-in technologies by schools, regions and on national levels. While the chock that came with a decline in Sweden’s PISA scores from 2006 on had haunted Swedish schooling, the new assessments and bio-datafication showed encouraging results that finally silenced even the most outspoken critics of school marketisation.

In Poodle’s new strategy, data on all aspects of students, from DNA profiling and precision-based adaptive instruction became valued as it provided a more complete picture of individuals and adjustments in feedback mechanisms to match their predicted study success. The models built around this data were used for predicting behaviour and these predictions were used to generalise to larger population demographics with surprisingly successful results. As schools, across the board from low to high achieving, gradually showed better and better results, politicians loudly announced that the Swedish school model worked and that the algorithmic performances of the big ‘learning machine’ platform worked too. Marketisation was described as finally closing bad schools, relieving parents of the need to choose between good and bad schools. Instead, school choice became a matter of a more superficial market choice, based on the rankings of student and teacher performances, choice of infrastructural add-ons, convenience and desired student or pedagogical profile.

Stage 3: adaptive platforms for an adaptive workforce

When Lina and Ingrid made the decision to move, they didn’t think much about schools. After all, the kids had done well at the local schools where they had lived before and they reasoned that ‘a school is a school’. Now, with the end of summer vacation approaching and the kids about to start the 2029 school year, a letter from their new school had caught them by surprise. The letter began as one might expect, welcoming them to the school community and informing them about the first day of school and the hours that the kids would attend. However, the third paragraph caught Lina and Ingrid off guard as it asked them to create a family account on Ninepenny-school if they didn’t already have one. ‘Ninepenny-school?’ said Ingrid. The kids had always gone to Poodle schools before and it hadn’t crossed their minds that the new school would be different. The paragraph continued, ‘If this is the first time your child has attended a school with the Ninepenny-school platform, please login to any previous school platforms, export your child’s data and import it to Ninepenny.’ Having heard some of the media discussion about growing Chinese influence in Sweden, Lina and Ingrid were somewhat reluctant to move their kids’ data, but chose to make the transfer. After all, both schools in the village they had moved to were using Ninepenny’s platform and they had decided that if there was a real problem, the government would have stepped in.

While the dominance of Poodle was seldom questioned, the middle 2020s saw a shift that did get a lot of media attention. Chinese investment in Swedish companies had been growing since the takeover of the car company Volvo at the end of the 2000s and Swedish subsidiaries to Chinese multinationals had quietly been buying up for-profit schools. In 2027, the world’s largest technology company, Ninepenny, took over these subsidiaries and promptly introduced its own digital platform. This led to almost a third of Swedish students receiving much of their instructional content, guidance and assessment from adaptive technologies based in Shenzhen rather than Mountain View.

Ninepenny logo.

While the prospect of an American platform being the dominant organising premise for Swedish schooling received critique at the local school level and from teacher unions who resisted what they understood to be a loss of control and professional status for teachers, the idea that a state controlled Chinese company could micromanage Swedish schooling quickly became an international political issue. There was fear that sensitive data about citizens would end up in the hands of a foreign state. Aided by European regulations, politicians responded to the perceived threat by restricting migration and mandating the storage of all school related data within the national borders. This development was a nuisance to Poodle who took the fight at both the national and European levels through its powerful lobby groups. While Poodle ultimately lost their battle leading to a drastic reduction in the commercial attractiveness of Swedish schooling for the company, Ninepenny stayed quiet throughout the process focussing instead on buying up schools and expanding their control over Swedish industry through their subsidiaries and partners. For them, data was not the goal unto itself but only a means to something bigger. For Ninepenny, where data was stored did not change the outcome.

Not all the differences between the two dominant platforms was not easily discernible on the surface, but for many teachers there were clear advantages to the new system. While some teachers had resisted Poodle’s platform on the grounds that it imposed a structure on lessons that they did not want to follow, Ninepenny’s platform was met with greater acceptance since it imposed far less structure and was perceived as more flexible. There were also major differences to be found in the adaptive learning systems and the kinds of instructional units’ students would engage in on the new platform that were a better fit with the types of group-based active learning pedagogies favoured by Swedish teachers. For Poodle, the individual had been the focal point of the system and the personalisation of content and precision-based adaptive learning had set the agenda, but Ninepenny’s algorithms worked on the group-level. Based on the system’s recommendations, students would be placed in groups of four and the entire learning process was now geared towards optimising the outcome of each group. For teachers, this was a welcome tool for orchestrating the classroom and individual learning at the same time. While the constellations could change during the years of compulsory education, teachers knew they would normally last for the entirety of their responsibility for a group during a three-year period of study. Even after graduation, it was not uncommon that such well fused units of learners would continue into higher education or working life together.

Teachers and stakeholders at a national level initially showed a great acceptance of the new system. Even once a degree of teacher resistance to what was seen by some as dictating students futures emerged, it was hard to argue with unemployment rates falling to a record low as more and more students seemed to find an interest in sectors where jobs were available. By optimising the local process of student feedback against goals articulated elsewhere, Ninepenny’s system began to shape interests and success at an individual level building on ideas articulated ten years earlier in a report funded by a newly Ninepenny acquired educational publishing company (Deegan and Martin 2018) and a book on ‘geno-economics’ (Plomin 2018). Coupling weighting of the neural networks used to provide feedback to the needs of various industries owned by the Ninepenny conglomerate and its partners, the schooling system began to generate a more ‘optimal’ workforce. A few critical voices were heard with a notable group of teachers forming what they described as a Humboldtian resistance intent on introducing holistic studies to school. Together with some critical academics who called for consideration of the public mission of education and claimed that the adaptive learning machines were a threat to democracy, the resistance argued that the Ninepenny systems were a threat to the integrity of individual humans and that they made people follow blindly. This echoed the warnings of Gert Biesta a decade earlier who characterised learning as something that, by its nature and narrow focus on the individual level, risks ‘keeping people in their place’ (2015, 7). However, while pockets of alternative schools did emerge, the complaints of naysayers were easily brushed aside in the mainstream political discourse. After all, the school system ranked the highest it ever had in international comparisons and unemployment was at a record low.

Coda

Based on a diagnosis of the current state of digital platform development, we have speculated and extrapolated on possible future scenarios for platformisation in the Swedish school system. These future scenarios that take place over the course of the 2020s form a speculative narrative in three stages. Stage one, schools and digital platform marketisation, describes how platform providers begin by selling platform services that service administrative tasks and then move into instructional content and assessment. In this way, they infrastructurally shape the organisation of schools and gradually become the de facto national school curriculum. Stage two, platform data exploitation, explains how platforms deliver essential services more efficiently and more cheaply to the state than earlier providers by profiting from user data. Global platform providers thereby constitute themselves as factual school. Stage three, adaptive platforms for an adaptive workforce, explicates how platform providers move to exploiting data by customised feedback loops at the group rather than individual level. Commercial profit is no longer the main purpose with the platform provided ‘for free’ while differentiating people to do certain things, fostering not only a global workforce but colonising data on human values and aspirations.

We argue that the three speculative future scenarios we have presented are a potential result of unintended consequences of policy and practice decisions that have already been made or may be made in the near future. This raises questions about how education platformisation affects public education values and national education systems, how it should be regulated and understood. In this final section, we will connect our speculative analyses to relevant critical conceptualisations of platformisation and discuss current considerations for researchers, educators and policy makers. In this discussion, we will focus on how we can make sense of the rise of digital platforms in schools and society more broadly, and what is made visible by focusing on a future historicising of digital platform development.

In our analyses of school platformisation, we have connected our speculations to what Srnicek (2017) argues is the continual developments of platform capitalism. From this perspective, the global platform industries can be seen to function first and foremost as economic actors that always seek business opportunity in the development of digital platforms. By historically contextualising digital platforms in relation to capitalism in society more broadly, Srnicek suggests that ‘phenomena that appear to be radical novelties may, in historical light, reveal themselves to be simple continuities’ (2017, 9). Our view is that this is very much the case with the platformisation of schooling with the accompanying introduction of data intensive practices and adaptive AI driven features, that while currently viewed as highly novel can be seen as part of a continuity that reaches back to the imaginary of Thorndike’s educational technology in the early 1900s. Srnicek (2017) proposes that four aspects characterise a platform. First, it has an intermediary function that enables interaction between different user groups and purposes. Second, a platform has an ability to extract data from users by constantly engaging users in platform activities. Third, it features cross-subsidisation, with the provision of free products and services meaning it accumulates more users and more activities on its network. Forth, a platform makes use of and thrives on network effects. In this sense, the larger a network gets, the more potential it has to extract and generate value from its users and their activities, and the faster it can expand and produce capital accumulation. Drawing on Srnicek’s conceptualisations, network effects are the main productive power in the speculative scenarios described here. With these scenarios, we argue that there is an inherent risk associated with an increase in personalised adaptive learning systems that may weaken or replace curricular values and the local characteristics of national education systems and cultures. The risk is that curricular values that have hitherto been democratically processed and negotiated may be replaced by de facto curricular values co-created by commercial interests and algorithmic powers.

In contrast with Keri Facer’s (2011) more optimistic speculations on an ideal of community-based schooling and education governance in the year 2035, we are more hesitant. She pictures individual ‘resource mapping’ as desirable to ‘provide a constant and ongoing record of everyone’s progress that’s much richer than it ever was in the past’ (122) in a school where mentors replace teachers. Facer also depicts a situation where local schools act as a concerned citizen, demanding government responsibility for knowledge resources, infrastructure and national information systems. Our deliberately pessimistic speculative scenarios raise the issue that the capacity to produce and control the infrastructures necessary for constant records and personalised adaptive systems may not inherently be something that governments have access to. Instead, the current landscape for the provision of digital infrastructure for public services suggests that those infrastructures are more likely to be controlled by multinational technology giants with their commercial and algorithmic powers expressing themselves in ways so intertwined with the everyday practices of schooling that they are difficult for governments to regulate.

Sweden as an example of a radically marketised and decentralised state governance system, provides a strong illustration of how local influence on how platformisation is made possible in the public education sector and can be at odds with the school as concerned citizen ideal that Facer argues for. It shows how when decision making about infrastructures is made at the local level, schooling may become particularly susceptible to the arguments of technology companies and to acceptance of commercial platform logics that allow global internet enterprises to reach into the very core of schooling. In our scenarios, a critical point is when a platform industry replaces the curricular processes of a democratic nation state such as constituting a shared purpose for public education and upholding an intellectual conversation on knowledge selection and choice of relevant teaching methods (Postman 2011). This speculative development stands in conflict with the common understanding that public education in democratic nations has as its main mission to foster new generations of citizens.

Attempts by commercial entities and other actors without a democratic mandate to influence education have often been framed in relation to this citizenship argument with the critic that citizenship is conceptualised as too narrowly market-oriented and neoliberal, resulting in curricula that only foster students for a (digital) labour market economy. However, as Biesta (2015) notes, citizenship-oriented curricula in the broadest sense are not only restricted to economic goals but also include existential and social aims that are more widely accepted. That is, that the purpose of schooling should not be to qualify citizens for the labour market, but to socialise them into society and local communities as well as to foster the ability to self-realise and be a subject of education. With the speculative scenarios presented here, our argument is that while curricula are under the control of nation states, their character and content with respect to the citizenship they engender may be subjected to the democratic processes of debate. However, with micro level control over the practices of schooling provided to commercial entities through their platforms, their influence and de facto curricula may sidestep the democratic processes associated with the governance of schooling entirely. Regardless of who the owner of a digital platform provider is or who it shifts to, a large global platform enterprise is not an instance of a globally acting and democratically responsible nation state government. It is not necessarily subject to the democratic processes generally associated with citizenship-oriented curricula models.

Another concern we wish to raise in relation to our speculative scenarios is the potential colonising powers of ‘datafied’ endeavours such as when digital platforms are introduced at geographically global and distributed scales and across socio-biologically levels. Taking up issues related to datafication in education and raising the potential of new forms of data colonialism, Couldry and Mejias (2018, 2–3), argue that the extraction and appropriation of human life through data, not only restricted to digital platforms but including wider forms of adaptations to datafication, ‘will provide the preconditions for a new stage of capitalism that as yet we can barely imagine’ (2). Their analysis suggests that similar to earlier forms of colonialism that profit on and appropriate territories, resources and people, the preconditions for data colonialism are fabricated as natural and rational. In this sense, personal data is naturalised as a resource that is ‘open for use’, ‘sharing’ is seen as the norm, and commercial extraction of such data sharing is seen as beneficial. This, Couldry and Mejias (2018) suggest, means that data or datafied social relations become the object of capitalist use while escaping the scrutiny that might otherwise be associated with such use. They reject the idea that open sharing of socially produced data is necessarily of societal benefit and instead raise the prospect that it as a,

commercially motivated form of extraction that advances particular economic and/or governance interests. Rejecting data colonialism does not mean rejecting data collection and use in all its forms. But it does mean rejecting the form of resource appropriation and accompanying social order that most contemporary data practice represents. (Couldry and Mejias 2018, 11)

Interestingly, in relation to the speculative scenarios we have presented here, Couldry and Mejias also refer to two dominant national contexts in the global flow of data, the United States and China. They relate the commercial dominance of these national contexts to circumstances that disrupt formerly entrenched East–West and North–South global distinctions. Our stage three speculative scenario draws on how an emerging Chinese state capitalist logic of platformisation could take hold of school infrastructure in other countries, opening up for a potential colonialism of this form of capitalist logic through micro level governance of the practices of schooling. Connecting Couldry and Mejias arguments about data colonialism specifically to education, our speculative scenarios also problematise the perspective that data openness is necessarily a public good. As Williamson (2017b) illustrates in his book, Big data in Education, educational data is a product sold and produced mainly for business interests, but circumscribed by seductive discourses of bright futures for students and societal prosperity. The digital governance inscribed into educational platform use and business logics is built upon the notion that data sharing, analysis and aggregation are inherently positive while it may also be aimed at maximising commercial interests such as profitability through further expansion and extraction of data in new spheres of human and social life.

In conclusion, we would like to pose the question of who owns and is responsible for the issues that the speculative scenarios presented here raise? Certainly, a clear argument can be made that national policy makers have a responsibility, but many of the issues we have raised here are related to decisions made by actors relatively distant from the spheres of national policy maker influence in countries like Sweden such as those responsible for the procurement of equipment and infrastructure in local schooling organisations. This includes the internet platform industry whose practices, as Macgilchrist (2019) has discussed, have started to occassion ethical and political problems. As issues emerge, the prevailing datafication regime inscribed into platforms that categorise, exclude and expose vulnerable populations like students is revealed.

To support both policy makers and those working more locally with decisions that may have unintended consequences like those we raise in our speculative scenarios, we argue that the efforts of researchers working in the area of the digitalisation of education are required. While there is already a large group of researchers examining the effects of digitalisation on the organisation of schooling and on learning processes, our open question to the community is, are we asking the right questions about how digital platforms are plugged into and prescribed to be used in the education sector? The tendency within the community to focus on the evidence-based question of ‘what works’ in schools may mask critical aspects important to raise in relation to the highly privatised and commercialised solutions of the educational technology industry (Courtney 2018, 28). Similarly, while clearly important, a focus on learning at the most local levels and on the development of students as individuals in relation to educational technology largely to the exclusion of a focus on more systemic effects may also mask potential problems. As platformisation moves into the educational sector bringing with it new forms of marketisation, the situation risks reformulating the broader mission of public education in ways unintended by both policy makers and educational practitioners. Particularly for countries with school systems already as decentralised and marketised as Sweden’s, these developments may bring with them significant issues. Issues that researchers working in the area of the digitalisation of education will need to consider, see early and respond to.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Thomas Hillman is Associate Professor of Education and researcher at the Department of Applied IT at the University of Gothenburg. Working at the intersection of the Learning Sciences, Human–computer Interaction, and Science and Technology Studies, he investigates the ways that technological change shapes learning and knowing.

Annika Bergviken Rensfeldt is Associate Professor of Education at the Department of Education, Communication and Learning at the University of Gothenburg. Based in the tradition of Sociology of Education, her interests include critical approaches to Educational Technology with a specific focus on policy and governance,and teachers’ digital work.

Jonas Ivarsson is Professor of Education and researcher at the Department of Applied IT at the University of Gothenburg. His interests concern the development of knowledge and technological change. Jonas’ research focuses on the role of representational technologies in the preservation, transmission and development of specialised knowledge and competence.

ORCID

Thomas Hillman http://orcid.org/0000-0003-3325-1204

Annika Bergviken Rensfeldt http://orcid.org/0000-0003-3936-2239

Jonas Ivarsson http://orcid.org/0000-0003-3402-525X