Teaching the Past to Build Equitable AI Futures with Emancipatory Artificial Intelligence

• In Event: Data Science Education: Re-Envisioning for 21st Century Teaching, Learning, and Research

Abstract

The Emancipatory Artificial Intelligence(EAI) framework(Monroe-White, forthcoming) equips data science (and beyond) researchers, educators, and policymakers with principles for fostering critical AI literacy. Rooted in critical quantitative(Zuberi & Bonilla-Silva, 2008) and critical computational approaches(Lee & Soep, 2016), EAI responds to Learning Sciences(LS) calls for integrating socio-historical context into technical computing education(Williamson & Eynon, 2020). Rather than treating algorithms and data as neutral—which LS scholars have warned are not(see (Morales-Navarro et al., 2024)—EAI highlights their entanglement with racial ideologies and systemic inequities(Monroe-White, 2021;Benjamin, 2019). By “unforgetting” harmful histories, learners begin to understand how AI can reinforce oppression or advance liberation(Vakil, 2018;Lee et al., 2021).

EAI pairs historical counter-archival storytelling(Milner & Howard, 2013) with open-inquiry data exploration(Acquah et al., 2024;Barany et al., 2024). This dual approach supports exploration of technical content(e.g., backpropagation, model design) and socio-historical contexts (e.g., Jim Crow, racial classification). Illustrative cases—algorithmic bias in face recognition (Buolamwini & Gebru, 2018), racialized predictive models(Angwin, 2016)—invite critique of statistical tools that long legitimized racial hierarchies(Galton, 1892;Pearson, 1901;Fisher, 1914) via community-based participatory inquiry(Tofel-Grehl et al., 2024;Rosenberg & Jones, 2024). The EAI foregrounds datasets and models, interrogating systems of power including why, how, and by whom they are collected, cleaned, and labeled so learners grasp how routine choices encode structural bias and why ethical data-science practice is pivotal for equitable AI outcomes.

EAI scales across professional learning, K-12, and higher education, offering design principles that blend technical skills with sociohistorical critique. Its three tenets—recognition, refusal, and repair—are designed to cultivate learner agency and critical inquiry. Recognition surfaces present harms such as algorithmic misogynoir(Noble, 2018;Bailey, 2021) and biased healthcare treatment(Obermeyer et al., 2019). Refusal mobilizes counter-archival narratives historicizing AI bias (e.g., Black women’s global hypervisibility and invisibility in science, media, and tech)(Noble, 2013;Mothoagae, 2016). Repair spotlights emancipatory exemplars like Du Bois who humanized Black communities through data visualization(Du Bois, 1899;Monroe-White & Lecy, 2023) and Joy Buolamwini, whose advocacy group, the Algorithmic Justice League, emerged from confronting racial bias in facial recognition. Hands-on work with contemporary tools lets students “peek under the hood” of classification systems while situating them within broader power relations(Lee, 2025;Walker & Schanzer, 2025;Weiland & Engledowl, 2022).

Motivated by the urgent need to train data scientists who can design, govern, and utilize AI in service of collective flourishing, EAI reinvisions what data to capture, whose stories to center, and which questions to ask of data-driven AI systems. EAI learners develop dual literacy: technical fluency with data and models that drive AI systems and critical awareness of their societal implications(Metcalf & Crawford, 2016;Connolly, 2010), Taken together, these capacities constitute re-envisioning the paradigm of data science education, whereby technical inquiry and civic imagination are equal contributors to advancing community well-being(Wright, 2020;Grover et al., 2024).

EAI reframes AI as a social science, revealing how power and equity shape and are shaped by data-driven AI innovations. It helps build learning environments grounded in historical accountability, community agency, and transformation(Monroe-White et al., 2025). Aligned with the theme, "Unforgetting Histories and Imagining Futures," EAI enables learners to holistically reflect and critique AI systems while embracing AI futures rooted in sovereignty and uplift.

Author

• Thema Monroe-White, George Mason University