“Literally Gold”: The Value and Vulnerability of Children’s Data in the Age of AI

• In Event: Highlighted Session: Capitalizing (on) a Global Learning Crisis: EdTech and the Political

Proposal

This paper examines the value and vulnerability of children’s data as technology companies compete to produce artificial intelligence (AI). While all data is valuable as companies train and refine their large language models and monetize it in different ways, children’s data, as a US security, technology, and education expert explained, is “literally gold.” This includes personally identifiable information and data on demographics, health, academics, behaviour, image, voice, and usage, much of which is collected by Big Tech and edtech companies (e.g. Williamson, 2017; Birch et. al., 2025). There is an increasing need to protect children’s data due to outdated data privacy regulations and security protections at national, state, and school levels (e.g. Hopper, et. al., 2022; 5Rights Foundation, 2025).

This paper is part of a broader study (2018-2025) on the influence of Silicon Valley venture capitalists and technology companies on education in the US and around the world. In this analysis, I explore three interrelated questions: 1) how and why is children’s data so valuable for AI and other technology companies; 2) how has the inadequate nature of data protection policies and security protocols left the data vulnerable; and 3) what (il)legal mechanisms are being used to access this data?

To explore these questions, the paper brings together Mejias & Couldry’s (2024) theory of “new data colonialism” and Moeller, et. al.’s (2024) educational capitalization to understand how children’s data functions as a frontier for AI profit. Mejias & Couldry (2024) call the phenomenon of unprecedented extraction of wealth from data the “new data colonialism,” defined as “a social order in which the continuous extraction of data from our lives generates massive wealth and inequality on a global scale” (p. 11). The extraction, valuation, and monetization of children’s data is therefore part of educational capitalization, what Moeller et. al. (2024) define as “the set of uneven processes and social relations through which value is extracted from educational processes and practices” (cite; Muniesa et. al., 2017). These processes occur through legal and illegal means, such as the hacking of PowerSchool, which serves 60 million students in over 90 countries (Merod, 2025).

To understand this phenomenon, I draw on a broader corpus of interviews, ethnographic observations, and documentary analysis that I have been collecting since 2018. The interviews for this paper are with US and UK technology and security experts, policymakers, school district digital learning specialists, venture capitalists, and education industry consultants. The observations have occurred at in-person and virtual industry events, such as three years of the ASU + GSV Summit. The documents include policy and legal documents, NGO reports, media articles, social media posts, and Crunchbase company investment records.

In conclusion, this paper will contribute to the broader literature on data platformization and surveillance in education by proposing a framework for mapping data extraction mechanisms (e.g. Singer, 2017; Williamson, 2017; Birch et. al., 2025; Feldman & Czerniewicz, 2023; Komljenovic, 2024), including those less discussed like mergers and acquisitions, school official designations, and hackings (e.g. Chapman, 2018; Richardson, 2020).

Author

• Kathryn Moeller