We live in the context of relentlessly ramping up galore around AI applied to education and education policy. Since the release of ChatGPT and the sensational rise in the use of chatbots combined with Large Language Models, AI has become a key discursive and material device for the government of education.

There is a need to explore how the rising centrality of AIED (Artificial Intelligence in Education) is producing changes in discourses and processes of education policy in European contexts. If the enmesh between public and private partnership in education has been abundantly problematised (Williamson,2017), in particular following the Covid-19 pandemic (Peruzzo et al,2022;Cone et al,2022), there is still the need to better understand how AI can be related to the emergence of a new set of political technologies for the government of education.

Here we focus on a particular aspect of those changes: the morphing dynamics between the state, the private sector and AI as a discourse and a constitutive part of the materiality of education policy and politics. We contribute to a more nuanced understanding of how AI becomes part of the education policy collective, who (or what) is recognised as a political actor and how the government of AI is articulated with, and conditions, the government of humans in education policy-making. The context we explore is that of the UK, where the new government has fuelled the investments in AI related facilities. The Prime Minister set out his plan to harness the potential of AI to drive growth and revolutionise British public services, including education.

By mobilising two intersecting discourses, AI as an equaliser of educational opportunities and outcomes, and saver of time for teachers cutting on routine practices, the government plans to develop prototype AI tools by April 2025 drawing on ’a first-of-its kind AI store data to ensure accuracy, so teachers can be confident in the information training tools’. £3m have been already devolved to promote the ‘Content Store Pilot’ led by Faculty AI. The aim is to create a bank of lesson plans and curricula to enhance the ability of AI generative tools to learn and ‘to add to the UK’s flourishing tech market, and support our mission to build sustained economic growth’(UK Gov,2024c).An additional £1m catalyst fund has been used to commission teaching and learning tools, built using the content store.16 technology developers were the fund, to ‘bring teachers and tech companies together to develop and use trustworthy AI tools’ (uk gov,2024a) and build trained workforce to deliver the skills needs of the next decade (Skills England,2024).

What is envisaged here is a hybrid governance modality combining state ownership of infrastructure, data and computing power (UK Gov,2025) with incentives aimed at private sector participation and investment. Drawing upon the merging of new materialist entanglements and Foucault’s governmentality (Lemke,2021) we use policy mobility and network ethnography (Ball,2012), to attempt to trace and problematise a whole set of relations, flows and meanings in UK education policy-making. To do so, we:

Map actors, events, funding, statements and the relationships between them;

Problematise such relations and the spaces and rationales they elicit,

Follow people and policies, visualising translations, re-contextualisation and making and re-making of AI education policy.

The paper intends to:

Shed light on the entanglements of actors and entities shaping policy-making contributing to a new political and educational reality where AI is an authoritative governing technology using the UK as an example of European country;

Show how AI as discourse and political materiality takes part in reallocating 'authority in relation to how education should be governed, by whom, to what ends and with what means’(Ball, 2016,13).

We use policy mobility as a methodological sensitivity to ‘focus on both multiple relationships inherent in the movement and connections of policies, people, and places and the various discursive and material flows these connections make possible’ (Gulson et al 2022,60). Our approach to policy mobility is informed by Lemke’s (2021) new materialist encounter with Foucault’s toolbox, mobilising dispositif and assemblages as arrangements of things, humans, scientific statements, money and subjectivities and extending the meaning of technology to apply it to ‘human affairs’ (p. 11). We look into: -The stirring policy movements Starmer’s plan has activated on different scales, in particular through the UK Sovereign AI body, which will fabricate, monitor and incentivise strategic AI collaborations and mobilities from local, national, and global companies and entrepreneurs; -The forms of governance that are being developed through the independent report AI Opportunities Action Plan (UK gov,2025a), the Generative AI in education Policy paper (UK gov, 2025b) -The role of the Science Advisory Council (uk gov,2024b), a team of (academic) experts appointed by the Department for Education to provide policy-makers with advice on strategic issues related to the delivery of higher standards of education, training and care. Methodologically, we use policy network ethnography to explore how different ways of doing policy and education governance are produced and animated and the kind of relationships that exist across them (Peck and Theodore 2015). The data were found on government and public websites, in news items, press releases, business biographies and other internet sources (podcasts, blogs) that contribute to a snapshot of policy activity. We created a database of organisations, programmes and people identified as network members through documents, and social media searches; and then identified and categorised these affiliations and ‘connections’, creating records onto the database, as the basis for the analysis of the work of networking. In analysing the data, we use Policy Network Analysis and the software Gephy to illustrate how these new actors, entities and fields are shaping the processes of policy-making by generating new policy claims, illustrating how these connections (networks) contribute to the changing nature of the politics of AI in education. Following Ball (2012,6), network is both ‘a conceptual device that is used to “represent a set of “real” changes in the forms of governance of education, both nationally and globally”’, and a method, ‘“an analytic technique for looking at the structure of policy communities and their social relationships’.

Our paper intends to give both a substantive and analytical contribution to the understanding of how AI can be related to the emergence of a new set of political technologies for the government of education and how it is involved into complex negotiations, enactments and stabilisations on the boundaries between the educational and the technological. We use a governmentality approach, sensitive to the government of things (Foucault, 2003; Lemke, 2021), to show the emergence of AI as both a discourse and material device in the dispositifs of education policy. We interpret the powerful positioning of the AI industry at the core of policy-making as foreshadowing a significant shift in contemporary neoliberal modes of government in education. Specifically, we see the AI-industry complex as playing a role in making possible and enhancing the management of fluctuating and unpredictable processes or, to put it differently, ‘governmental practices that seek to steer and manage performances and circulations by acting on and controlling the heterogeneities and differences that make up a milieu’ (Lemke, 2021, 126), what Foucault calls environmentality (Foucault, 2003). Both at the level of discourse and contextualisation, AI open the space for thinking and enacting practices that set in motion the capacities of a nonhuman entity to recalibrate dysfunctional educational systems and security mechanisms that seek to anticipate future emergencies. At the same time, the articulation of the policy networks we have mapped, captures how state actors act as initiator or facilitator of the private sector participation in education policy but are, at the same time, faced with a new space of unknows directions, possibilities and threats which call for new dispositifs to governing radically heterogenous things, AI in our case.

Ball, S. J. (2012). Global Education Inc. New policy networks and the neo-liberal imaginary. London and New York: Routledge. Ball, S. J. (2016). Ball, Stephen J. (2016) Neoliberal Education? Confronting the Slouching Beast. Policy Futures in Education, 14(8), 1046–1049. Cone, L., K. Brøgger, M. Berghmans, M. Decuypere, A. Förschler, E. Grimaldi, S. Hartong, et al. 2021. “Pandemic Acceleration: Covid-19 and the Emergency Digitalization of European Education.” European Educational Research Journal, OnlineFirst, 1–24. doi:10.1177/14749041211041793. Foucault, M. (2003). The Birth of Biopolitics: Lectures at the Collège de France, 1978-1979. Palgrave Macmillan. Gulson, K. N.; Sellar, S, and Webb, P.T. (2022). Algorithms of Education: How Datafication and Artificial Intelligence Shape Policy. Minneapolis, London: University of Minnesota Lemke, T. (2021). The Government of Things. Foucault and the New Materialisms. New York: New York university Press Peck, J., & Theodore, N. (2015). Fast Policy. Experimental Statecraft at the Thresholds of Neoliberalism. University of Minnesota Press. Peruzzo, F.; Ball, S. J. and Grimaldi, E. (2022). Peopling the crowded education state: heterarchical spaces, EdTech markets and new modes of governing during the Covid-19 pandemic. International Journal of Education Research, https://doi.org/10.1016/j.ijer.2022.102006 Rabinow, P. (2008). Marking Time: On the Anthropology of the Contemporary. Princeton, N.J.: Princeton University Press Selwyn, N. (2022). The future of AI and education: some cautionary notes. European Journal of Education, 57, 620-631, DOI: 10.1111/ejed.12532 UK governament, (2024a) https://www.gov.uk/government/news/teachers-to-get-more-trustworthy-ai-tech-as-generative-tools-learn-from-new-bank-of-lesson-plans-and-curriculums-helping-them-mark-homework-and-save) UK government (2024b) https://www.gov.uk/government/news/department-for-education-establishes-science-advisory-council UK Government, (2024c), Minister Morgan GEIS speech https://www.gov.uk/government/speeches/minister-morgans-geis-speech UK government, 2025a https://www.gov.uk/government/publications/ai-opportunities-action-plan/ai-opportunities-action-plan UK government, 2025b, https://www.gov.uk/government/publications/generative-artificial-intelligence-in-education/generative-artificial-intelligence-ai-in-education) Williamson, B. (2017). Williamson, B. (2017). Big Data in Education: The Digital Future of Learning, Policy and Practice. London: Sage. Williamson, B. and Hogan, A. (2020). Williamson, B., & Hogan, A. (2020). Commercialisation and Privatisation in/of Education in the Context of Covid-19. Bruxelles: Education International.