Content uploaded by Ben Williamson
Author content
All content in this area was uploaded by Ben Williamson on Dec 08, 2022
Content may be subject to copyright.
1
Governing through infrastructural control:
artificial intelligence and cloud computing
in the data-intensive state
Ben Williamson, University of Edinburgh
Author manuscript of chapter published in Housley, W., Edwards, A., Beneito-Montagut, R. &
Fitzgerald, R. (eds). 2022. The Sage Handbook of Digital Society. London: Sage.
https://us.sagepub.com/en-us/nam/the-sage-handbook-of-digital-society/book269045
Introduction
Digital data have become integral to the ways societies, institutions, and citizens are monitored,
understood, and governed. Numerical information has played a long historical role in how states
and governments monitor and manage their territories and citizens (Foucault, 2007). The
twentieth-century rise of computing technologies, the internet, and data science has opened up
the task of governing to more diverse actors and new modes of quantification, manipulation and
control in the twenty-first century (Edwards et al, 2009). Since around 2010, attention has turned
to the expanding governance capacities of computation and ‘big data’ analytics, algorithms,
machine learning and artificial intelligence (Johns, 2021), with businesses, governments and non-
state actors increasingly treating digital data as a ‘major object of economic, political and social
investment for governing’ human lives, institutions, industries and societies (Bigo, Isin &
Ruppert, 2019, p. 2).
Data, however, do not appear without an apparatus of data production, processing, and
dissemination. Underpinning all data are the necessary infrastructures to generate, analyze and put
them to use, although the work performed by these systems often remains hidden, only
becoming of concern when they break down or fail to operate as intended (Bowker & Star,
1999). More specifically, digital data infrastructures are complex networked systems consisting of
computational and database technologies that underpin an increasing array of daily activities and
processes across diverse sectors worldwide (Kitchin, 2014). Data infrastructures are ‘comprised
of shifting relations of databases, software, standards, classification systems, procedures,
committees, processes, coordinates, user interface components and many other elements which
are involved in the making and use of data’ (Gray, Gerlitz & Bounegru, 2018, p. 3). They include
the global infrastructures of technology firms that facilitate information access and social
interaction online; urban data infrastructures that support city services; intelligence
infrastructures for national and international security; financial data infrastructures for market
transactions and trading; cloud infrastructures that host and power third-party platforms and
online services; and sector-specific data infrastructures on which services such as health, justice,
welfare and education depend, as well as many others (Edwards et al, 2013; van Dijck, 2013;
Easterling, 2014; Plantin & Punathambekar, 2019).
Data infrastructures are not merely technical backdrops to practices of information retrieval,
sociality, urban life, public service delivery, and so on, but have integrated into how societies
2
function at various scales and in myriad ways (van Dijck, Poell and de Waal, 2018; Prainsack,
2019). The ways authorities govern depends on digital infrastructures of services, data
processing, and the production of insights and knowledge from streams of data about citizens,
events, spaces, and objects, which may then be used for decision-making and other governance
actions (Isin & Ruppert 2019; Fourcade & Gordon, 2021). The data generated from such
infrastructures also carry huge financial value, as digital information is treated as an asset and
source of economic and societal value creation (Birch, Cochrane & Ward, 2021).
The diversity of digital data infrastructures means understanding how they function, and their
consequences, requires attention to the contexts of their production and deployment. Education
is one such context, with both the schools and higher education sectors becoming increasingly
‘datafied’ (Jarke & Breiter, 2019; Williamson, Bayne & Shay, 2020). The datafication of education
is an evolution of historical modes of ‘governing by numbers’, and the role of statistics and
metrics in reshaping the institutions, practices, purposes and values of education (Biesta, 2009;
Piattoeva & Boden, 2020). A range of digital data infrastructures of varying scales and
functionality are already active in education (Fenwick, Mangez & Ozga, 2014). These include
national, state-level and international infrastructures of large-scale testing and performance
measurement; digital infrastructures such as learning management systems and student
information systems; knowledge and research management infrastructures; educational
technology (edtech) infrastructures for digital teaching and learning; and global cloud computing
infrastructures such as Amazon Web Services, Google Cloud, and Microsoft Azure, which host
many of the digital services and platforms that schools, colleges, and universities depend upon
for pedagogical, curricular, assessment, management, data management and reporting purposes
(Williamson, 2017).
Viewed as a historical series, educational data infrastructures are evolving from those of large-
scale statistical measurement to new approaches using big data, cloud computing, machine
learning, predictive analytics and AI that operate at fast pace, expanding scope, and ‘intimate’
levels of granularity (Beneito-Montagut, 2017; Gorur, 2018; Perrotta & Selwyn, 2020; Williamson
& Eynon, 2020). Examining the ‘ongoing infrastructuring of educational governance’ through
database technologies and networks (Ratner & Gad, 2019, p. 541), and more recently cloud
computing and machine learning, also opens up wider questions about the role of data
infrastructures in other core areas of government action, particularly how they are ‘reconfiguring
the work and organization of states’ (Johns, 2021, p. 5).
The chapter first includes a general conceptual section on governing infrastructures. It then
surveys the various infrastructural arrangements underpinning recent developments in education
by the global technology companies Amazon, Google and Microsoft, as ‘big tech’ firms with
colossal techno-economic power (Birch & Cochrane, 2021). It focuses particularly on how these
global technology corporations are building new governing infrastructures through cloud
computing and machine learning. The final section outlines how digital data infrastructures
signify evolving modes of governance and power, and shifts in the ways that human populations
and individual lives may be traced, acted on, and governed by both states and corporations.
Governing infrastructures
Numerical data collection and use has been integral to the tasks of governments since at least the
early nineteenth century, as states deployed the new sciences of statistics, demography and
probability to enumerate and manage their territories, estates and citizens (Hacking, 1982; Rose,
1999; Ambrose 2015), and to describe ‘the reality of the state itself’ (Foucault, 2007, p. 274).
3
Following the growth of personal computing, the worldwide web, and computational storage in
the late twentieth century, statistical governance has gradually mutated into novel forms of
governing based on the use of data analytics, algorithms and artificial intelligence or machine
learning for processing heterogeneous data streams at huge scale and rapid pace (Isin & Ruppert
2020). The ‘technocratic infrastructure’ of ‘database government’ represents a shift from the
‘qualitative’ governance of the social to the ‘quantitative’ governance of the ‘informational,’
where governance is increasingly achieved through scraping information traces from societies,
social relations and individual lives (Ruppert, 2012, pp. 117-118).
Just as statistical governance depended on techniques and instruments of quantification such as
surveys, censuses, graphical displays and reports, governance by data also involves a complex
apparatus of data collection and analysis instruments and practices. Data infrastructures
encompass the various social and technical aspects involved in the collection, connection,
calculation, communication, and consumption of data, including all the ‘the institutional,
physical, and digital means for storing, sharing and consuming data across networked
technologies’ (Kitchin, 2014, p. 32). These networked social, technical, economic and political
elements are rarely stable, as other programs, applications, actors and software may be ‘plugged
in’, thereby expanding the infrastructure, changing its form, and reworking the possibilities of its
use and effects (Plantin et al, 2018). Therefore, a data infrastructure is ‘not so much a single thing
as a bundle of heterogeneous things’, and is always ‘braided with social, cultural and political
actors and their values’ which together function to generate new ‘social facts’ (Slota & Bowker,
2017, p. 531-32). The politics of data infrastructures include the preferences, knowledges, values
and practices of those who built and programmed the infrastructure and which shape how it can
be used and to what purposes. Data infrastructures are not separable from the power, knowledge
and expert assumptions of their planners or producers, the labour required for their functioning,
repair, and maintenance, their relations with other systems and infrastructures, and the
ideological work involved in imagining, assembling, and maintaining them (Plantin &
Punathambekar 2019).
Digital data infrastructures have changed how corporations and states can govern, steer and
control their objects and human subjects. According to Johns (2021, p. 3), ‘governance by data’
denotes ‘the propensity for the gathering, assemblage, formatting, analysis, transmission, and
storage of digital data to have governance effects’. This form of data-infrastructural governance,
which involves dispersed and distributed multisector actors rather than state agencies alone,
generates new knowledge and categories for understanding human lives, events, institutions and
societies and leads to novel data-driven interventions (Isin & Ruppert, 2020). The progressive
penetration of digital technologies and corporations in the functioning of state institutions and
public sector services is part of a ‘transformation in political rationality, in which data
affordances increasingly drive policy strategies’ (Fourcade & Gordon, 2020, p. 78). Like other
infrastructures of transport, communication and power, data infrastructures are the shared
accomplishment of ‘new constellations of international, inter-governmental and
nongovernmental players’ and constitute a form of distributed, multisector ‘extrastatecraft’
(Easterling, 2014, p. 15).
Increasingly, the forms of extrastatecraft enacted by digital data infrastructures are being
extended and advanced in their capacities by the addition of machine learning and artificial
intelligence, as infrastructures become ‘hubs of command and control’ across myriad sectors,
industries and services (Bigo et al, 2019, p. 13). This involves automated systems and data-driven
decision-making in key state agencies and functions (Calo & Citron, 2021), including the
4
emergence of AI as state actors (Crawford & Schultz, 2019), cloud computing and machine
learning (Amoore, 2020), and ‘thinking infrastructures’ that combine human and machine
cognition in the governance of states and markets (Kornberger et al, 2019). Fourcade and
Gordon (2021, p. 86) describe the emergence of a ‘dataist state’ that governs by modelling its
population as a set of fluctuating and emergent patterns, perceives social problems in terms of
individual rather than collective causes, is oriented to the near future through predictions based
on data, and manages through behavioural manipulation.
In this context, public institutions such as government, health, welfare and education ‘have
become predicated on the use of private online infrastructures’ for many of their functions,
including through the hosting of new public service platforms on private infrastructures (van
Dijck et al, 2018, p. 16). Education is exemplary of how private infrastructures have been made
integral to state and public systems. As in other domains of governance, the use of data in
educational governance has a long history, initially in official government agencies’ use of
statistics for system-level monitoring (Lawn, 2013). The introduction of New Public
Management (NPM) as a paradigm of public sector reform in the 1980s/90s demanded
increasing quantitative performance measurement for accountability purposes, as materialized in
the expansion of large-scale testing and school comparison; teacher performance evaluations; the
rise of research and teaching evaluations in universities; league tables, rankings and other means
of ‘scoring’ and ‘rating’ individuals and institutions; and mechanisms for acting on them in the
name of performance ‘improvement’ (Grek, Maroy & Verger 2021). Previous research on data
infrastructures in education has focused on the systems required for processing test data, such as
those of state agencies, education and testing businesses, and international large-scale assessment
agencies, and their enactment of political agendas of performance monitoring and accountability
(Anagnostopoulos, Rutledge & Jacobsen, 2013; Sellar 2015; Lewis and Hartong, 2021).
More recently the task of educational governance has been distributed to a wider array of actors
and digital technologies with their own social and technical capacities. New digital infrastructures
in schools and universities extend beyond test data to applications such as mass student
monitoring, educational big data mining, learning analytics, and other digitalized measures of
teaching, learning and institutional effectiveness and improvement (Gorur, Sellar & Steiner-
Khamsi, 2019; Wyatt-Smith, Lingard & Heck, 2021). Educational technologies and digital
platforms such as learning management systems, online learning platforms, and a host of other
applications, often linked together in interoperable networks, are also becoming infrastructural to
the functions and actions of schools, colleges and universities, and governing what forms of
education, pedagogy, curriculum and assessment can take place (Decuypere, Grimaldi & Landri,
2021).
Emerging scholarship has begun to examine the creation of digital data infrastructures of cloud
computing and artificial intelligence, as predictive data analytics, adaptive platforms, and machine
learning have been developed for ‘real-time’ assessment of students, automated decision-making,
and the ‘personalization’ of learning (Williamson & Eynon, 2020). These developments signify a
‘shift of responsibility from the teacher, the school leader or the bureaucrat to the algorithm, the
new automated manager who makes decisions for organisations and institutions’ (Souto-Otero &
Beneito-Montagut, 2016, p. 22). Gulson and Witzenberger (2021, p. 1) refer to ‘automated
education governance’ as ‘system-level and school-level practices that are exercised by automated
decision-making machines, and instances in which software has a role in governing education’.
Automated governance operates through AI such as machine learning, which uses data to model
and predict school performance and students’ cognition, knowledge, comprehension (Perrotta,
5
2021), and even affective states such as ‘student mental health’ or ‘non-cognitive skills’ (Kotouza
et al, 2021).
Like other data infrastructures, educational data infrastructures consist of heterogeneous
networks of people, organizations, technologies and policies that stretch across and beyond the
boundaries of formal education systems to include education companies, software firms,
consultancies, foundations, and research centres, as well as formal education agencies
(Anagnostopoulos et al, 2013; Hartong 2018). Educational data infrastructures evolve, mutate,
are made interoperable with other platforms and programs, and are put to new tasks and
purposes either by design or incrementally as technical innovations emerge and new business
cases for their use are established (Gulson & Sellar, 2019). Educational data infrastructures also
encode particular desires, values, and morals, defining through their evaluative criteria what
‘counts’ as a ‘good school’, a ‘good teacher’ or a ‘good student’ (Sellar, 2015). Beyond being
complex, relational, sociotechnical systems, data infrastructures and other connected platforms
and applications also act as sociotechnical relays of policy objectives to reform and transform
education (Williamson, 2019).
The ongoing evolution of contemporary data infrastructures of education to include capacities of
analytics, machine learning, and the cloud is therefore part of a political program involving a
variety of actors from multisectoral positions in a radical ‘reimagining’ of education in schools,
colleges and universities. The chapter next turns to surveying the interrelated infrastructural
arrangements enabling cloud computing and AI to become a governing power in education, with
specific attention to the unfolding competition for infrastructural dominance and control in
education between Amazon, Microsoft and Google.
Surveying infrastructural arrangements
Infrastructure imaginaries
Data infrastructures are products of specific networks of actors with particular objectives, desires
and visions of the future. Jasanoff (2015, p. 4) refers to ‘sociotechnical imaginaries’ as
‘collectively held, institutionally stabilized, and publicly performed visions of desirable futures’.
They act as powerful aspirational and normative visions of society and technology, which often
originate in particular social settings and gain force as they become seemingly consensual and
even common sense articulations of desirable pathways for both technical development and
social progress. Similarly, the cognate concept of ‘data imaginaries’ captures the visionary
‘promises, futures, potentials and possibilities that are associated with data’ by governments and
industry (Beer, 2019, p. 9). These imaginaries pave the way for data infrastructure development
and enactment. They propose a future vision of society that can be attained through
technological development, such as the construction and operationalization of the digital data
infrastructures that are integral to advancing new forms and applications of cloud computing,
data analytics, machine learning, and AI. Organizations including Google, Amazon and
Microsoft are powerful producers and promoters of sociotechnical imaginaries of social
transformation through cloud computing and machine learning (Amoore, 2020).
In the context of education, imaginaries supportive of data infrastructure development have
been produced and circulated by the wider social and political networks that constitute the
sector. Since around 2010, an increasing number of powerful organizations, operating across
sectors and at scales from the regional and national to international contexts, has coalesced
around a shared future imaginary in which education will be powered, augmented, enhanced and
improved by data analytics, machine learning and automation (Williamson & Eynon, 2020).
6
These include accelerated ‘real time’ measurement rather than reliance on discrete assessments
and evaluations conducted at long temporal intervals; automated responses, feedback,
‘personalized learning’ adaptations; and even ‘pre-emptive’ interventions generated by AI itself
(Witzenberger & Gulson, 2021). As such, imaginaries of AI infrastructures reflect both a history
of increasing datafication of education and a future vision in which automatization, real-time
insight generation through analytics, and the adaptivity of AI are seen as desirable and attainable.
While educational governance by numbers historically rested on an apparatus of measurement
instruments and expert quantitative techniques sited in formal governing agencies (Piattoeva &
Boden, 2020), its imagined evolution through the cloud and AI relies for its attainability and
enactment on digital data infrastructures built by commercial technology companies and
supported by influential organizations including the OECD, World Bank and World Economic
Forum (Williamson, 2020).
Google, Microsoft and Amazon have become key organizations in the production and
dissemination of powerful future imaginaries of education, positioning themselves and their
evolving technologies as transformative infrastructures for pedagogy, assessment and
management. Microsoft, illustratively, published a detailed report on AI in education, entitled
‘Transform Student Engagement: Achieve Personalized, Efficient, Inclusive, and Accessible
Higher Education with AI’, which focused on ‘AI-powered hyperpersonalized, accessible, and
inclusive learning experiences and tools’:
Digital technologies such as artificial intelligence are transforming higher education, just like any
other sector. … AI is being combined with … intelligent task/process automation to support
innovative offerings like digital portals/applications for ‘out of classroom’ interactions,
personalized learning, virtual classrooms, and virtual tours. Soon, institutions will be able to
provide each student with his or her own AI-driven digital assistant to act as an ever-present
learning assistant and personal tutor. (Jyoti & Sutherland, 2020, p. 2)
Similarly, AWS proposes ‘artificial intelligence and machine learning can help higher education’
through ‘on-demand, tailored content’ for ‘personalized learning’ (AWS Public Sector Team,
2019), while Google has promoted AI add-ons for its cloud-based educational services (Perrotta
et al, 2021).
Imaginaries are powerful catalysts of the infrastructural developments required to introduce the
cloud and AI into educational settings and contexts. Cloud and AI imaginaries animate and
invest in specific forms of technical development while simultaneously producing political
conviction in a shared vision of digital transformation. Through the production, stabilization and
circulation of imaginaries, new market opportunities have opened up for global private
infrastructure companies such as Google, Amazon and Microsoft to participate in state and
public education systems. Understanding that requires a shift in analytical attention from
animating imaginaries to the organizational strategies underpinning the materialization of
infrastructures of cloud computing and AI in education.
Organizational strategies
A digital data infrastructure is always tightly related to an organizational strategy and business
model. Governance through data infrastructure is therefore inseparable from wider
organizational dynamics and their underpinning imaginaries. The corporate competition for
market power in cloud computing and AI is one such dynamic, as Google, Amazon and
Microsoft in particular have sought to extend their share of the ‘cloud infrastructure and
platform services’ market (Bala et al, 2021). Understanding the roles of Google Cloud, Amazon
7
Web Services and Microsoft Azure as creators and vendors of cloud infrastructures used to
govern diverse industries and sectors of society therefore requires scrutiny of the business model
of the cloud (Narayan, 2022).
Cloud computing emerged from visions in industrial and academic research and development of
computing as a ‘utility’ and ‘service’, and has remained since around 2010 a fast-paced site of
technical and business innovation (Varghese & Buyya, 2017). Influentially defined by the US
National Institute of Standards and Technology (NIST), cloud computing consists of three
service models for consumers: software as a service (SaaS), platform as a service (PaaS), and
infrastructure as a service (IaaS). NIST also identifies five characteristics of all cloud services: on-
demand self-service, broad network access, resource pooling, rapid elasticity, and measured
service; and highlights the cloud computing model as ‘a more economic method of providing
higher quality and faster services at a lower cost to the users’ (Liu et al, 2011, p. 1).
Though cloud computing has a long developmental history in computer science, the
contemporary techno-economic business model of cloud computing appeared with the full
launch of Amazon Web Services (AWS) in 2006 to enable other businesses to access
computational services hosted in its large data centre facilities on demand (Varghese et al, 2019).
As such, cloud computing is both a field of innovation in computer systems and a new business
model aimed at delivering efficiencies and cost-savings to customers while centralizing
computing resources in a small number of technology corporations (Weinhardt et al, 2009). The
increasing dependency of global technology corporations such as Amazon, Microsoft and
Google on cloud computing ‘for their financial success means that the companies use political,
economic and technical resources to ensure that the clouds are the “default” infrastructure in as
many domains as possible’ (Fiebig et al, 2021, p. 11). By the 2020s, the driving business model
behind these ‘big tech’ businesses had created ‘a new multi-sided ecosystem’ from which they
could ‘demand both a toll and masses of data’ (Birch & Cochrane, 2021, p. 5).
AWS was originally envisaged as a ‘backbone’ for other companies to run their business from the
cloud (Gartenberg, 2021). Built on two key services, the Elastic Compute Cloud (EC2) and
Simple Storage Service (S3), AWS constituted a new techno-economic business model, making it
the leader in public cloud computing with more than 30% global market share ahead of main
competitors Microsoft Azure and Google Cloud. It has also expanded beyond corporate
customers to government, state and public sector operations. As an ‘architecture of market
power’, AWS serves as essential infrastructure for other businesses and government services that
now depend upon it (Khan, 2017, p. 710). Microsoft Azure was officially launched in 2010 as an
Infrastructure-as-a-Service and Platform-as a-Service competitor to AWS. Its services included
Azure Blob storage, compute, network and database services, Azure Data Lake Store and Azure
Data Lake Analytics to offer big data and analytics, and machine learning Workbench and ML
Services for AI functionality and integration (Janakiram, 2020). Like AWS, Microsoft has sought
to expand Azure into new markets and sectors, including health, government and education.
Similarly, Google Cloud provides a suite of storage, compute, network, database, analytics and
AI services (Google Cloud Content Team 2021). The Google Workspace suite of productivity
and enterprise tools and platforms are embedded in Cloud, and Google signs agreements with
partner organizations to extend the functionality of the infrastructure and its integrated services.
It also partners extensively with public sector organizations, facilitated by a specific Google
Cloud Public Sector channel, especially in government services and education.
8
Amazon, Microsoft, and Google have all incrementally developed their cloud business models to
become the backbone of a vast range industries, sectors and services. The underpinning business
model of elasticity and scalability, twinned with the technical expandability of cloud
infrastructures, has allowed them to extend in an unprecedented way into sectors such as
government, health and education, while generating huge profit advantages. The business model
conforms to an emerging techno-economic paradigm of ‘assetization’ under ‘technoscientific
capitalism’, whereby value can be generated from the ownership and control of something as a
long-term revenue stream rather than traded as commodities for immediate financial gain (Birch,
2020). Digital data infrastructures such as AWS, Azure and Google Cloud are key assets for their
corporate owners and controllers, generating value through subscriptions, multitenancy rents,
and on-demand fees, as well as from the data they collect and process (Birch & Cochrane, 2021).
Within these infrastructures, user data are captured, repurposed, and converted into assets too—
rendered valuable through the storage, aggregation and analysis facilities provided by the cloud—
as ‘users’ become techno-economic objects for monetization and value-creation as future
revenue streams (Birch et al, 2021). In this respect, the assetization of data as a key aspect of
contemporary technoscientific capitalism is interdependent with its digital cloud and data
infrastructures (Narayan, 2022).
Moreover, as the techno-economic business model of these corporations is built-in to the
sociotechnical data infrastructures they construct, the model also leaves its imprint on the sectors
and services it touches. As ‘architectures of market power’, AWS, MS Azure and Google Cloud
seek infrastructural dominance across sectors, actively changing the sectors they penetrate to
conform to their imaginaries and strategies (Khan, 2017). Within the field of education, cloud
infrastructures inscribe the techno-economic logics of assetization on to institutions and
practices of teaching and learning, as institutions pay subscriptions and rents for access to
storage, computing and analytics services from the key cloud rentiers, and those cloud operators
derive value from the user data generated from the integration of their infrastructures with
educational institutions (Komljenovic, 2021). Those techno-economic strategies materialize in
the ways cloud architectures perform managerial and pedagogic functions.
Cloud architectures
Global cloud infrastructures have become increasingly tightly integrated into education as their
corporate proprietors have sought to advance their cloud assets and secure economic rents from
customers and subscribers across an expanding range of sectors and industries, especially since
the Covid-19 pandemic resulted in widespread uptake of online learning platforms and services
(Komljenovic, 2020). Universities in the US and UK in particular have become dependent on
corporate cloud infrastructures for many of their core digital services and functions since around
2015:
Instead of running IT services with on-site teams and on infrastructure owned by organizations,
services are now often deployed on public cloud infrastructure. … However, this operational
paradigm shift also leads to a change in control. While, before, user data would remain on
infrastructure controlled by an organization, this data is now stored and processed by an external
operator. (Fiebig et al, 2021, p. 1)
Cloud architectures have become integral to many aspects of the three core activities of
universities: research, teaching and administration. The reliance on cloud services in education
raises issues of ‘infrastructure and data control’ (ibid), as institutions’ abilities to audit or
implement privacy, ensure data protection compliance, or obtain meaningful informed consent
for data collection and use are limited and constrained when they employ cloud operators.
9
Moreover, in relation to the teaching function, clouds introduce new capacities of data analysis
and algorithmic intervention that may shape and steer educators’ practices and students’ learning
experiences. Clouds exert control over curricula and govern the ‘means of study’ (ibid., p. 7) by
determining what kinds of educational tasks are possible to implement.
As the three dominant global vendors of cloud infrastructures for education, Microsoft Azure,
Google Cloud, and AWS provide a vast range of services to education institutions. Azure, for
example, includes ‘virtual desktop’ and ‘app virtualization that runs in the cloud’ to support
online and remote teaching and learning, ‘automatic remote management and scaling’, and
‘Azure Lab Services’ and ‘virtual machines’ for running classes and managing research, plus
‘flexible migration’ to ‘the cloud with managed service’
(https://www.microsoft.com/mea/azureforedu/). Google Cloud for education provides
‘scalable infrastructure’ for educational institutions, including its Workspace suite of tools
(formerly known as G Suite) for digital learning, ‘virtual desktop infrastructure’, ‘data warehouse
optimization’, ‘serverless apps for edtech’, ‘smart analytics and AI’ and ‘intelligent learning tutors
powered by AI’ to support ‘personalized learning’
(https://cloud.google.com/solutions/education). AWS for education also offers ‘Desktop-as-a-
Service’ solutions and ‘personal cloud desktops’ for distance learning, ‘no-cost online learning
modules on cloud computing’, and ‘virtualized app streaming’ for ‘anytime access’, as well as full
institutional migration to the AWS cloud (https://aws.amazon.com/education/).
One key feature of the digital data infrastructures of cloud computing in education is the
promise to create ‘data lakes’ and harness machine learning algorithms to derive insights from
very large reservoirs of heterogeneous data (Perrotta, 2021). According to AWS, ‘The three steps
to gaining the full value of data are: migrate data to the cloud data lake, set up an analytics
engine, and leverage artificial intelligence (AI) technologies, including ML and deep learning’
(https://aws.amazon.com/blogs/publicsector/the-strategic-power-of-data-enabled-by-aws-
partner-data-led-migrations/). The process of ‘architecting a data lake’ involves the deployment
of multiple AWS products and functionalities, including those for pulling student learning data
from all integrated datasets, and then utilizing several AWS programs for handling the ‘machine
learning workload’ of analysis (Jordan & Berkley 2020). The data lake is the source from which
machine learning algorithms can distil and condense insights, as, ‘Using the AWS Cloud, schools
and districts can get a comprehensive picture of student performance by connecting products
and services so they seamlessly share data across platforms’
(https://aws.amazon.com/education/K12-primary-ed/). AWS also promotes ‘Machine Learning
for Education’ services to ‘identify at-risk students and target interventions’, ‘improve teacher
efficiency and impact with personalized content and AI-enabled teaching assistants and tutors’,
and ‘improve efficiency of assessments and grading’ (https://aws.amazon.com/education/ml-in-
education/).
Following the imaginary and business strategies outlined earlier, these technology corporations
are seeking to become fully infrastructural substrates to educational institutions. Each of them
promises to facilitate institutions’ digital transformation and to introduce capacities for enhanced
digital teaching, including the increased gathering of large quantities of student and institutional
data in data lakes, and the use of the data analytics and AI capacities that cloud infrastructures
provide. As such, the creation of digital infrastructures in education ‘relies increasingly on
engineering “openings” for the seamless insertion of external big data and predictive
technologies’ which ‘are appealing for their promise of computational efficiency, but also opaque
10
and unknowable for their dependencies that extend deep into the broader ecosystems of Google
and Amazon’ (Perrotta, 2021, p. 68).
The migration of education to data lakes, cloud infrastructures and their algorithmic facilities
opens up new ways of measuring, understanding, and governing both individuals and
institutions. Amoore (2020) approaches cloud computing in terms of its algorithmic powers of
perception. The cloud, she argues, ‘is a bundle of experimental algorithmic techniques acting in
and through data’, and ‘contemporary cloud computing is about rendering perceptible and
actionable (almost seeing) that which would be beyond the threshold of human vision’ (ibid, p.
41). When an algorithm in ‘cloud space’ determines an action, it calculates plural possible
pathways and narrows them down by discarding much of the material it has encountered. By
extracting patterns, norms and anomalies from vast data collections, cloud computing condenses
the complexity of the volume of digital traces to single outputs and identified targets of
intervention. Crucially, as apparatuses that make targets perceptible for governing interventions,
cloud computing arrives at its targets through complex chains of correlation, inference and
prediction, such that its identification of targets might never have been possible through
conventional techniques of observation and identification. Thus the outputs of the data lake and
the cloud are not ‘objective’ representations of the world, but ‘generative’ of particular
configurations of possible future states of the world, which then ‘open onto targets of
opportunity, commercial and governmental’ (ibid, p. 43).
In the case of education, such ‘targets’ might include a student whose anticipated future
academic progress or personal well-being is determined to require pre-emptive action in the
present (Witzenberger & Gulson 2021). The promise of cloud architectures of data lakes,
predictive analytics, and automated pre-emptive interventions is to configure a new kind of cloud
classroom or campus with advanced, automated perceptive powers and capacities to generate
‘actionable’ insights and targets of attention. The cloud campus and classroom engages in
comprehensive monitoring not only to catalogue events as they happen but to model and
recognize future states, and introduce automated, pre-emptive actions to govern behaviours and
operationalize actions towards optimal goals (Andrejevic, 2019). The evolution of data
infrastructures to incorporate cloud-based machine learning and AI means they operate as
‘thinking infrastructures’ of distributed human–machine cognition, which can structure attention
and orchestrate decision making, influence thought and action, and enable new forms of
governance, power and control (Kornberger et al, 2019). Within the field of education, cloud-
based, cognitive, thinking infrastructures enable new forms of ‘automated thinking’ that make
targets of intervention perceptible, shape human decision making, and create new possibilities
for educational policymaking and the governance of school practices (Sellar & Gulson, 2021).
These powers also extend through ‘plug-and-play’ integrations with third-party platforms.
Platform integrations
A key aspect of digital infrastructures is the development of ‘gateway’ technologies and
agreements which allow different systems to interoperate as if they were a single integrated
system and ‘to form more powerful and far-reaching networks’ (Edwards et al, 2009, p. 369).
Two key developments need to be considered here: the rise of ‘platforms’ that integrate into
digital data infrastructures, and ‘application programming interfaces’ (APIs) that provide the
gateways for platforms, infrastructures and other platforms to interoperate (Plantin et al, 2018;
van Dijck et al, 2018). Digital platforms are ‘applications whose technical architecture
emphasizes the provision of connection, programmability, and data exchange with applications
developed by others’ (Plantin et al, 2018, p. 296). APIs govern the rules through which such
11
platforms may be integrated and made interoperable with computational infrastructures and
other platforms, acting as ‘key infrastructural elements and connectors’ which create ‘a
standardized grammar of functionality … that typically helps to make them interoperable and
independent of their respective implementations’ (Snodgrass & Soon 2019, n.p.) APIs also
facilitate the delegation of computational tasks such as data storage and processing tasks to cloud
and analytics infrastructures.
APIs allow Google, Microsoft and Amazon to create more far-reaching and powerful networks
in education based on interoperability standards between their cloud infrastructures and third-
party platform integrations. They have created interoperability protocols for third-party
platforms to integrate with their clouds, such that their infrastructures are becoming not only
integral to institutions’ own systems, but act as a substrate to a much wider ecosystem of third-
party educational technologies and platforms too. Beyond their capture of institutional
information infrastructure, the global cloud operators also act as essential infrastructure for other
education technology operators, who depend on their storage and computing facilities for the
functioning of their educational products and services. This reflects how global technology firms
augment their techno-economic ecosystems through integrating ‘outsiders’ with their
infrastructures and capitalizing on both the rents paid and expanded data extraction capacities
that result from their interoperation (Birch & Cochrane, 2021).
Google Classroom, for example, is an integrated platform of the Google Workspace for
Education suite of cloud-based platforms and tools. Originally launched in 2014 as Google’s
flagship online learning platform, by 2021 Classroom had a reported reach to over 150 million
students worldwide. Behind the platform is a specific Classroom API allowing third-party
platforms to integrate with Classroom (Perrotta et al, 2021). Google has actively leveraged the
API to create a Marketplace of ‘Classroom add-ons’, to enable thousands of third-party
integrations, enabling it to introduce new data analytics features by drawing down user data from
the full interoperable Classroom and Workspace markets. Classroom is a platform intermediary
between schools and the Google Cloud, enabling schools to access cloud-powered services while
Google, in return, can gather valuable data from its users for further product refinement, feature
upgrades and other developments. It positions the Google Cloud as infrastructural to a vast
segment of the edtech industry, as third-party vendors must adapt and agree to the rules of the
API to interoperate with the Classroom.
Similarly, AWS employs a specific API as a gateway to educational institutions, in the shape of its
Alexa Education Skills API for either institutions to create their own voice interfaces or for
edtech companies to build voice features for existing products. AWS is also infrastructural to the
edtech industry by providing cloud service support to both startups and established education
companies. For startups, its AWS EdStart ‘accelerator’ program is ‘designed to help
entrepreneurs build the next generation of online learning, analytics, and campus management
solutions on the AWS Cloud’ (https://aws.amazon.com/education/edstart/). For the edtech
industry more generally, AWS also claims to be ‘accelerating transformation in educational
technology’ through its cloud services:
The Amazon Web Services (AWS) Cloud enables education technology (EdTech) companies to
accelerate development of scalable and secure technology solutions that support students and
educators every day. AWS helps EdTechs realize the full potential of cloud computing with
dedicated business and technical resources to support growth.
(https://aws.amazon.com/education/ed-tech/)
12
AWS is fully integrated with some of the globe’s largest and most highly capitalized edtech
companies and education businesses, including the international learning management systems
Canvas, Blackboard and Schoology, and the multinational massive online learning platform
Coursera. AWS enables Coursera to collect data about students and facilitates its analysis, which
Coursera mobilizes as sources of insight for further product refinement and development to
enhance engagement and uptake at large scales.
Through their myriad platform integrations, Google, Microsoft and AWS are seeking to make
themselves seemingly essential to education. Their ambitions are typical of large-scale
infrastructure program managers’ desires to ‘scale up’:
Behind every system builder’s ambition stands the hope of network effects, in which utility
increases exponentially with the number of users: a system with thousands of users might be
worthwhile, but a system with millions of users is an industry, and one with billions of users—like
the global telephone network or the Internet itself—becomes obligatory. (Edwards et al, 2009, p.
370)
AWS, Microsoft and Google education all reach millions of users, and through their ever-
widening platform integrations they are becoming virtually obligatory infrastructures on which
both the edtech industry and educational institutions have come to depend for digital access,
computation and data processing. The network effects of these infrastructures and their platform
integrations mean AWS, Google and Microsoft are able to assetize user data at significant scale,
in accordance with their overarching business models and imaginaries. However, as with all
infrastructures, the potential for these clouds to become obligatory architectures for education
also depends on their penetration into working practices.
Infrastructure work
Digital data infrastructures demand new forms of work and labour, in terms of their production
and maintenance as well as their embedded uses in specific contexts. In the context of
proliferating data work in contemporary societies, Burrell and Fourcade (2021, p. 5) argue that a
‘new occupational class’ of computational experts has become authoritative across a range of
industries and sectors, which they term the ‘coding elite’. The coding elite comprises technology
entrepreneurs, software engineers, computer science academics, and data science professionals,
among others, who advance computation, software code, algorithms and data analytics as the key
to transformations in industry, government, and other sectors:
In its quest for market expansion, the tech industry increasingly carves away at and lays claim to
tasks that once were protected as the proper domain of professional judgement in every
occupation from business management, medicine, and the criminal justice system to national
defense, education, and social welfare. … Legitimacy has been displaced from the professional to
the coder-king—and, increasingly, to the algorithm. (ibid, p. 6)
Since the early 2010s, a new coding elite has entered education too, in the shape of
‘computational experts, data-savvy administrative staff and large corporations inserting
themselves through back-ends and various other dependencies’ (Perrotta, 2021, p. 69). This elite
consists of experts in ‘education data science’, ‘learning analytics’, ‘AI in education’, and ‘learning
engineering’ (Williamson, 2020), as well as other new data infrastructure professionals who are
required to maintain and monitor systems, take responsibility for data generation and analysis,
and mobilize the outputs in decision-making processes (Lewis & Hartong, 2021). Schools,
colleges and universities have to recruit new positions for data officers, system administrators
and other technical roles (Selwyn, 2021), all of whose ‘data practices’ enable a digital data
13
infrastructure to function as intended and produce effects (Decuypere, 2021). These new data
experts in schools, colleges and universities perform significant roles in making and analyzing
student data, and using insights derived from the quantification of learning activities and
outcomes to inform decisions for managerial and pedagogic purposes (Whitman, 2020).
The expansion of Google, Amazon and Microsoft infrastructures in education is introducing
new capacities of an emerging ‘cloud elite’ into the ways universities, colleges, and schools
function and are governed. All three feature dedicated education teams and partnership
programs to promote use of cloud infrastructures in education, and support the use of machine
learning, data analytics and AI to transform existing practices, procedures and cultures of
education. AWS, in particular, has established Cloud Innovation Centers with universities around
the world to ‘test new ideas with Amazon’s innovation process, and access the technology
expertise of AWS’ (https://aws.amazon.com/government-education/cloud-innovation-
centers/). Google, Microsoft and Amazon all run extensive programs to train and certify
educators in their products, and work with ‘broker’ organizations who support institutions to set
up, run and maintain their infrastructures and platforms. These infrastructure habituation
programs are intended to govern how educational institutions, staff and students use the cloud,
promising the significant benefits of cloud-based teaching and learning, data storage and
analytics to schools, colleges and universities while simultaneously seeking to attract students to
future career pathways in cloud computing (Perrotta, 2021).
The new cloud elite of computational and data analytics experts is introducing new capacities
into schools, colleges and universities to govern their pedagogic, administrative and management
functions. It is also creating new stratifications of knowledge and expertise, prioritizing technical
and data experts who can both work with the cloud to generate insights, and then narrate and
translate the results into decision-making and other ‘actionable’ interventions. In this way,
‘apparently global systems’ are translated into actual use in concrete contexts and ‘penetrate
deeply into the work lives of organizations and individuals’ (Edwards et al, 2009, p. 370).
In addition to the skilled cloud labour of data infrastructures are the data subjects whose
academic work—teaching and studying—is recorded in the cloud for analysis and potential
action. While the ‘coding elite’ has consolidated power through their technical control over the
digital means of data production, this also depends on extracting data from an unpaid
‘cybertariat’ that is ‘yoked into the day-to-day operations of algorithmic systems’ (Burrell &
Fourcade, 2021, p. 7). Teachers become cybertariat labour for cloud data infrastructures, while
students become the source of the freely extracted data these vast systems depend on for both
their financial valuation and core functions of data analysis, machine learning and AI. The
algorithmic analysis of data by cloud infrastructures creates ‘datafied’ students and educators, as
‘a student’s identity is increasingly constructed for them through the analysis of “their” data by a
number of usually unseen third parties—such as platform providers, data brokers, businesses,
government, public institutions and the state’ (Selwyn, Pangrazio & Cumbo, 2021, p. 3).
The cloud conjures up specific configurations of student and educator subjects, who are
identified, measured and classified through the in-built criteria of the infrastructure as
‘algorithmic identities’ (Cheney-Lippold, 2011) or ‘digital proxies’ (Johns, 2021). These data
selves are not stable identities but called up on demand through the technical practices of linking
relational databases, creating data lakes, and employing the sorting, classifying and predictive
capacities of machine learning (Burrell & Fourcade, 2021). In the educational context, as Pickup
(2021) argue, these data practices ‘invent’ student identities, rendering them as digital proxies or
14
‘informational people’ who are known in bits and bytes and thereby made amenable to
evaluation, prediction, and potential targeted intervention. In these data-intensive ways,
‘educational selves’ become so inextricably tied to data ‘that students and teachers become
unidentifiable without it’ (ibid, p. 8).
In these ways, digital data infrastructures both configure new forms of expert work, thereby
privileging particular computational and analytical elites, and create new understandings of
human subjects that may be consequential in terms of decision-making or action derived from
data analysis. ‘Determination of subjects’ relative noteworthiness or value for governance
purposes is increasingly guided by digital sensing and processing of which those subjects have no
awareness—processes of determination often entangled with commercial activities and
incentives’ (Johns, 2021, p. 8). New positions are opened up by the construction of digital data
infrastructures for private technology companies to become governing experts, in education as
well as other sectors, and to exert influence over the ways organizations operate. Professional
positions for certified cloud educators, infrastructure specialists, data analysts and other technical
posts distribute the governing logics of data infrastructures through the labour of a widening
array of actors and activities. Microsoft, Amazon and Google-certified educators and
administrators are trained and habituated to work with the cloud according to the corporate
imaginaries and business models of these global corporations, and act as relays of the cloud into
everyday practices. Through such dynamics, students and educators are reconstituted as
informational proxies, and institutions as datafied schools and cloud campuses from which value
may be constantly extracted. Moreover, they signify distinctive shifts in how states and
companies seek to govern.
New states of infrastructural power and governance
In this chapter digital data infrastructures have been examined as ‘distributed accomplishments,
constituted by an evolving set of relationships between people and devices, software and
standards, words and instruments’, and ‘which afford their own ways of knowing and
possibilities for action’ (Gray et al, 2018, pp. 9-10). Focusing specifically on education, it has
shown how commercial cloud infrastructures are increasingly central to the measurement and
management of schools and universities, reflecting the historical evolution of ‘governance by
numbers’ in education (Piattoeva & Boden, 2020), and charted the emergence of experimental,
data-intensive decision-making and governance (Sellar & Gulson 2021). Infrastructures of test-
based performance measurement in education have begun to evolve into more artificially-
intelligent information networks hosted in the cloud and powered by machine learning
algorithms, which promise new capacities of prediction, pre-emption, and ‘automated
governance’ interventions (Witzenberger & Gulson, 2021).
The chapter has surveyed the sociotechnical imaginaries, techno-economic strategies, cloud
architectures and machine learning operations, platform integrations and interoperabilities, and
the varied forms of labour and identification involved in infrastructural forms of governance in
education. Focusing on Google, Amazon and Microsoft, it has shown how global big tech firms
are expanding their cloud computing and artificial intelligence capacities into schools, colleges
and universities, in ways that make educational institutions, educators and staff amenable to new
forms of datafication, automation and control. They are also competing for cloud share in
education, as part of their techno-economic business plans for infrastructural dominance in as
many sectors and industries as possible.
15
The implications of this competition for infrastructural control extend beyond education. Data
infrastructures are recomposing a vast range of other state activities, including national statistics,
public welfare provision, criminal justice administration, intelligence and policing, environmental
protection, humanitarian relief, emergency management, and more (Johns, 2021). Automated
systems, predictive analytics and artificial intelligence are increasingly performing as state and
public service actors (Calo & Citron, 2021). This is bringing private technology companies into
the performance of core state functions, with ‘techniques and logics of state governance’ shifting
to ‘streams of digital data for purposes of trying to understand the needs, wants, and
circumstances of their constituents’ and thereby ‘giving rise to new ways of seeing and being as a
state’ (Johns, 2021, p. 5).
New developments in data infrastructures introduce novel forms of governance and behavioural
control into state and public sector operations (Yeung, 2016). These infrastructural evolutions in
the constitution and function of the state are emblematic of the appearance of a particular form
of governance that Isin and Ruppert (2020) term ‘sensory power’. Sensory power relies on
specific configurations of technologies, expert knowledge, practices, and agencies, authorities
and other organisations, which are able to ‘make people sense-able’ from automated surveillance
of their digital traces. Automated surveillance and comprehensive monitoring through
distributed, embedded, always-on sensing networks shifts logics of governance towards
prediction, pre-emption, and ‘simulated futures’ as the basis for ongoing interventions in the
present (Andrejevic 2019). Instead of ‘periodic stocktaking’, sensory power derives from tracing
live ‘signals’ and, via machine learning analysis, making visible the ‘pulses, flows and patterns’ of
individuals, then assigning them to associated clusters and categories (Isin & Ruppert, 2020, p.
10). It then involves acting back and intervening recursively, in the pursuit of improving their
performance.
The forms of educational re-infrastructuring by cloud corporations documented in this chapter
exemplify the emergence of sensory power as a new mode of public-private state governance
enabled by digital data infrastructures. Sensory power is part of a new form of ‘digital statecraft’
in an ‘artificially intelligent state’ that tends to ‘learn’ about its citizens from their digital data
traces and defer decision-making to machine learning algorithms, and in which ‘the displacement
of analog technologies by digital ones, as well as emergent forms of computation, are defining
new affordances and possibilities for state bureaucracies to see, sense, and act’ (Fourcade &
Gordon, 2020, pp. 84-85). Fourcade and Gordon (2020, p. 82) argue that government and public
sector institutions in health, welfare, justice and education often lack the material resources or
technical expertise to gather or analyse data, and therefore increasingly outsource their
‘administrative agency decisions to artificially intelligent systems’ provided by corporate
infrastructure companies. This ‘corporate reconstruction of the state’ is evident in digital firms’
attempts to access and capitalize on data produced by the state, ultimately enabling increasingly
‘state-like corporations’ to ‘reinvent traditional public functions’ (ibid. p. 78).
Amazon, Google and Microsoft, as their infrastructural incursions into education demonstrate,
now occupy the position of state-like corporations, with social, technical, economic and visionary
power to shape how institutions, individuals, and whole systems are measured, evaluated, rated,
predicted, and controlled and governed. The evolution of corporate digital data infrastructures to
incorporate cloud computing, machine learning algorithms and AI is introducing new modes of
sensory power and data-driven digital statecraft into core areas of government and the public
sector such as education. Rather than a statistical state that sees and governs its citizens and
subjects through broad brush metrics captured at long intervals, the artificially intelligent state
16
delegates power to ‘state-like’ corporate technology companies and cloud infrastructures that can
‘sense’, make perceptible, and govern subjects ‘live’ and in ‘real time’ through data analytics and
machine learning.
References
Ambrose, M. 2015. Lessons from the avalanche of numbers: big data in historical perspective. I/S: A
Journal of Law and Policy for the Information Society 11(2): 201-277.
Amoore, L. 2020. Cloud Ethics: Algorithms and the attributes of ourselves and others. London: Duke University
Press.
Anagnostopoulos, D., Rutledge, S.A. & Jacobsen, R. (eds). 2013. The Infrastructure of Accountability: Data use
and the transformation of American education, 1-20. Cambridge, MA: Harvard Education Press.
Andrejevic, M. 2019. Automating surveillance. Surveillance and Society 17(1-2), 7-13.
AWS Public Sector Team. 2019. How artificial intelligence and machine learning can help higher
education. AWS Public Sector Blog, 19 December: https://aws.amazon.com/blogs/publicsector/how-
artificial-intelligence-and-machine-learning-can-help-higher-education/
Bala, R., Gill, B., Smith, D, Ji, K and Wright, D. 2021. Magic Quadrant for Cloud Infrastructure and
Platform Services. Gartner, 27 July: https://www.gartner.com/doc/reprints?id=1-
26YXE86I&ct=210729&st=sb
Beer, D. 2019. The Data Gaze: Capitalism, power and perception. London: Sage.
Beneito-Montagut, R. 2017. Big data and educational research. In: Wyse, D., Smith, E., Suter, L.E. and Selwyn,
N. (eds.) The BERA/Sage Handbook of Educational Research, 913-933. London: Sage.
Biesta, G. 2009. Good education in an age of measurement: on the need to reconnect with the question
of purpose in education. Educational Assessment, Evaluation and Accountability 21, 33–46
Bigo, D., Isin, E. & Ruppert, E. (eds). 2019. Data Politics: Worlds, subjects, rights. Abingdon: Routledge.
Birch, K. 2020. Technoscience Rent: Toward a Theory of Rentiership for Technoscientific Capitalism.
Science, Technology and Human Values 45(1), 3-33.
Birch, K. & Cochrane, D.T. 2021. Big Tech: Four Emerging Forms of Digital Rentiership. Science as
Culture: https://doi.org/10.1080/09505431.2021.1932794
Birch, K., Cochrane, D.T. and Ward, C. 2021. Data as asset? The measurement, governance, and
valuation of digital personal data by Big Tech. Big Data and Society:
https://doi.org/10.1177/20539517211017308
Bowker, G.C. & Star, S.L. 1999. Sorting things out: Classification and its consequences. London: MIT press.
Kornberger, M., Bowker, G.C., Elyachar, J., Mennicken, A., Miller, P., Randa Nucho, J. and Pollock, N.
(eds), 2019. Thinking Infrastructures. Bingley: Emerald Publishing.
Burrell, J. & Fourcade, M. 2021. The society of algorithms. Annual Review of Sociology 47, 23.1–23.25.
Calo, R. & Citron, D.K. 2021. The Automated Administrative State: A Crisis of Legitimacy. Emory Law
Journal 70(4), 797-845.
Cheney-Lippold, J. 2011. A new algorithmic identity: soft biopolitics and the modulation of control,
Theory, Culture & Society 28(6), 164-81.
Crawford, K. & Schultz, J. 2019. AI systems as state actors. Columbia Law Review 119(7), 1941–72
17
Decuypere, M. 2021. The Topologies of Data Practices: A Methodological Introduction. Journal of New
Approaches in Educational Research 10(1): https://doi.org/10.7821/naer.2021.1.650
Decuypere, M., Grimaldi, E. and Landri, P. 2021. Introduction: Critical studies of digital education
platforms. Critical Studies in Education 62(1): 1-16.
Easterling, K., 2014. Extrastatecraft: The power of infrastructure space. London: Verso.
Edwards, P.N., Bowker, G.C., Jackson S.J. & Williams, R. 2009. Introduction: An agenda for
infrastructure studies. Journal of the Association for Information Systems 10(5), 364–374
Edwards, P.N., Jackson, S.J., Chalmers, M.K., Bowker, G.C., Borgman, C.L., Ribes, D., Burton, M., &
Calvert, S. 2013. Knowledge Infrastructures: Intellectual Frameworks and Research Challenges. Ann Arbor: Deep
Blue.
Fenwick, T. Mangez, E. & Ozga, J. (eds). 2014. Governing Knowledge: Comparison, Knowledge-Based Technologies
and Expertise in the Regulation of Education. London: Routledge.
Fiebig, T., Gürses, S., Gañán, C.H., Kotkamp, E., Kuipers, F., Lindorfer, M., Prisse, M. & Sari, T. 2021.
Heads in the Clouds: Measuring the Implications of Universities Migrating to Public Clouds. arXiv
preprint: arXiv:2104.09462.
Foucault, M. 2007. Security, Territory and Population: Lectures at the College de France 1977-1978. Trans. G.
Burchell. New York: Palgrave Macmillan.
Fourcade, M. & Gordon, J. 2020. Learning Like a State: Statecraft in the Digital Age. Journal of Law and
Political Economy 1(1), 78-108.
Google Cloud Content Team. 2021. Cloud Covered: What was new in May on Google Cloud. The
Keyword, 18 June: https://blog.google/products/google-cloud/cloud-covered-what-was-new-may-
google-cloud-2021/
Gorur, R. 2018. Escaping numbers? Intimate accounting, informed publics and the uncertain assemblages
of authority and non-authority. Science & Technology Studies 31(4), 89–108.
Gorur, R., Sellar, S. & Steiner-Khamsi, G. (eds). 2019. Comparative Methodology in the Era of Big Data and
Global Networks. Abingdon: Routledge.
Gray, J., Gerlitz, C. & Bounegru, L. 2018. Data infrastructure literacy. Big Data & Society:
https://doi.org/10.1177/2053951718786316
Grek, S., Maroy, C. & Verger, A. (eds). 2021. Accountability and Datafication in the Governance of Education.
Abingdon: Routledge.
Gulson, K.N. & Sellar, S. 2019. Emerging data infrastructures and the new topologies of education policy.
Environment and Planning D: Society and Space 37(2), 350–366.
Gulson, K.N. & Witzenberger, K. 2022. Repackaging authority: artificial intelligence, automated
governance and education trade shows. Journal of Education Policy 37(1), 145-160.
Hacking, I. 1982. Biopower and the avalanche of printed numbers. Humanities in Society 5, 279-295
Hartong, S. 2018. Towards a topological re-assemblage of education policy? Observing the
implementation of performance data infrastructures and ‘centers of calculation’ in Germany.
Globalisation, Societies and Education 16(1), 134–150.
Isin, E. & Ruppert, E. (2020). The Birth of Sensory Power: how a pandemic made it visible. Big Data and
Society, July-December, 1-15.
18
Janakiram, M.S.V. 2020. A Look Back At Ten Years of Microsoft Azure. Forbes, 3 February:
https://www.forbes.com/sites/janakirammsv/2020/02/03/a-look-back-at-ten-years-of-microsoft-
azure/
Jarke, J. & Breiter, A. 2019. Editorial: the datafication of education. Learning, Media and Technology, 44(1),
1–6.
Jasanoff, S. 2015. Future imperfect: Science, Technology, and the Imaginations of Modernity. In S.
Jasanoff & S.-H. Kim (eds). Dreamscapes of Modernity: Sociotechnical Imaginaries and the Fabrication of Power, 1-
33. Chicago: University of Chicago Press.
Johns, F. 2021. Governance by data. Annual Review of Law and Social Science 17, 4.1–4.19.
Jordan, C. & Berkley, T. 2020. Architecting a Data Lake for Higher Education Student Analytics. AWS
Architecture blog, 22 October: https://aws.amazon.com/blogs/architecture/architecting-data-lake-for-
higher-education-student-analytics/
Jyoti, R. & Sutherland, H. 2020, April. Transform Student Engagement: Achieve Personalized, Efficient,
Inclusive, and Accessible Higher Education with AI. IDC:
https://edudownloads.azureedge.net/msdownloads/IDC-Transform-Student-Engagement.pdf
Khan, L.M. 2017. Amazon’s Antitrust Paradox. The Yale Law Journal 126, 710-805.
Kitchin, R. 2014. The Data Revolution: big data, open data, data infrastructures and their consequences. London:
Sage.
Komljenovic, J. 2020. The future of value in digitalised higher education: why data privacy should not be
our biggest concern. Higher Education. https://doi.org/10.1007/s10734-020-00639-7
Komljenovic, J. 2021. The rise of education rentiers: digital platforms, digital data and rents. Learning,
Media and Technology: https://doi.org/10.1080/17439884.2021.1891422
Kotouza, D., Callard, F., Garnett, P. & Rocha, L. 2021. Mapping mental health and the UK university
sector: Networks, markets, data. Critical Social Policy: https://doi.org/10.1177/02610183211024820
Lewis, S. & Hartong, S. 2021. New shadow professionals and infrastructures around the datafied school:
Topological thinking as an analytical device. European Educational Research Journal:
https://doi.org/10.1177/14749041211007496
Narayan, D. 2022 Platform capitalism and cloud infrastructure: Theorizing a hyper-scalable computing
regime. Environment and Planning A: Economy and Space 54(5), 911-929.
Perrotta, C. 2021. Programming the platform university: Learning analytics and predictive infrastructures
in higher education. Research in Education 109(1), 53–71.
Perrotta, C. & Selwyn, N. 2019. Deep learning goes to school: Toward a relational understanding of AI in
education. Learning, Media and Technology 45(3), 251-269.
Perrotta, C., Gulson, K.N., Williamson, B. & Witzenberger, K. 2021. Automation, APIs and the
distributed labour of platform pedagogies in Google Classroom. Critical Studies in Education 62(1), 97-
113.
Piattoeva, N. & Boden, R. 2020. Escaping numbers? The ambiguities of the governance of education
through data. International Studies in Sociology of Education 29(1-2), 1-18.
Pickup, A. 2021. Toward a historical ontology of the infopolitics of data-driven decision-making
(DDDM) in education. Educational Philosophy and Theory:
https://doi.org/10.1080/00131857.2021.1935232
19
Plantin, J.-C. & Punathambekar, A. 2019. Digital media infrastructures: pipes, platforms, and politics.
Media, Culture & Society 41(2), 163-174.
Plantin, J.-C., Lagoze, C., Edwards, P.N. & Sandvig, C. 2018. Infrastructure studies meet platform studies
in the age of Google and Facebook. New Media & Society 20(1), 293-310.
Prainsack, B. 2020. The political economy of digital data: Introduction to the special issue. Policy Studies
41(5), 439–446.
Ratner, H., & Gad, C. 2019. Data Warehousing Organization: Infrastructural Experimentation with
Educational Governance. Organization 26(4), 537–552.
Rose, N. 1999. Powers of Freedom: Reframing political thought. Cambridge: Cambridge University Press.
Ruppert, E. 2012. The Governmental Topologies of Database Devices. Theory, Culture & Society 29(4-5),
116-136.
Sellar, S. 2015. Data infrastructure: a review of expanding accountability systems and large-scale
assessments in education. Discourse: Studies in the Cultural Politics of Education 36(5): 765-777.
Sellar, S. & Gulson, K.N. 2021. Becoming information centric: the emergence of new cognitive
infrastructures in education policy. Journal of Education Policy, 36(3), 309-326.
Selwyn, N. 2021. The human labour of school data: exploring the production of digital data in schools.
Oxford Review of Education 47(3), 353-368
Selwyn, N., Pangrazio, L. & Cumbo, B. 2021. Knowing the (datafied) student: the production of the
student subject through school data. British Journal of Educational Studies:
https://doi.org/10.1080/00071005.2021.1925085
Slota, S.C. & Bowker, G.C. 2017. How infrastructures matter. In: Felt, U., Fouche, R., Miller, C.A. &
Smith-Doerr, L. (eds). The Handbook of Science and Technology Studies, 4th ed, 529-54. London: MIT Press.
Snodgrass, E., & Soon, W. 2019. API practices and paradigms : Exploring the protocological parameters
of APIs as key facilitators of sociotechnical forms of exchange. First Monday, 24(2).
doi:10.5210/fm.v24i2.9553
Souto-Otero, M. & Boneito-Montagut, R. 2016. From governing through data to governmentality
through data: artefacts, strategies and the digital turn. European Educational Research Journal 15(1), 14-33.
van Dijck, J. 2013. The Culture of Connectivity: A Critical History of Social Media. Oxford: Oxford University
Press
van Dijck, J., Poell, T. & de Waal, M. 2018. The Platform Society: Public values in a connective world. Oxford:
Oxford University Press.
Varghese, B. & Buyya R. 2017. Next generation cloud computing: New trends and research directions.
Future Generation Computing Systems 79(3), 849-861.
Varghese, B. et al. 2019. Cloud Futurology. Computer 52(9), 68-77.
Weinhardt, C. et al 2009. Cloud Computing – A Classification, Business Models, and Research Directions.
Business Information Systems Engineering 1, 391-399.
Whitman, M. 2020. ‘We called that a behavior’: The making of institutional data. Big Data and Society:
https://doi.org/10.1177/2053951720932200
Williamson, B. 2017. Big Data in Education: The digital future of learning, policy and practice. London: Sage.
20
Williamson, B. 2019. Policy networks, performance metrics and platform markets: charting the expanding
data infrastructure of higher education. British Journal of Educational Technology 50(6): 2794-2809.
Williamson, B. 2020. New digital laboratories of experimental knowledge production: artificial intelligence
and education research. London Review of Education 18(2): 209-220.
Williamson, B., Bayne, S., & Shay, S. 2020. The datafication of teaching in higher education: critical issues
and perspectives. Teaching in Higher Education 25(4): 351-365.
Williamson, B. & Eynon, R. 2020. Historical threads, missing links, and future trajectories in AI in
education. Learning, Media & Technology 45(3): 223-235.
Witzenberger, K. & Gulson, K.N. 2021. Why EdTech is always right: students, data and machines in pre-
emptive configurations. Learning, Media and Technology: https://doi.org/10.1080/17439884.2021.1913181
Wyatt-Smith, C., Lingard, B. and Heck, E. (eds). 2021. Digital Disruption in Teaching and Testing: Assessments,
big data, and the transformation of schooling. Abingdon: Routledge.
Yeung, K. 2016. Hypernudge: big data as a mode of regulation by design. Information, Communication and
Society 20, no. 1: 118-136.
