1. Topic, Research Questions, and Objectives

By 2026, Generative Artificial Intelligence (GenAI) has transitioned from a supplementary tool to foundational "cognitive infrastructure". However, a phenomenon termed the "Double-Speed Classroom" is emerging: a crisis where premium algorithmic resources are cordoned off behind a structural economic barrier. This research interrogates the “$20 Intelligence Barrier": a term derived from the de facto industry standard pricing (e.g., ChatGPT Plus, Claude Pro) which imposes a fixed monthly cost of $20 USD globally, regardless of local purchasing power.

This research addresses two primary questions:

1. How does this flat-rate "Cognitive Tariff" create quantifiable stratification between students in different socio-economic strata?

2. How can institutional leadership transition from "passive recipients" of technology to "institutional architects" capable of neutralizing these inequities? The objective is to formalize and validate the LEGA framework, providing administrators with a rigorous diagnostic toolkit to identify and mitigate systemic algorithmic inequity.

2. Conceptual and Theoretical Framework

This study integrates Bourdieu’s theory of Cultural Capital with Jan van Dijk’s "Third-level Digital Divide," which focuses on the inequity of outcomes resulting from the quality of resource access. We also draw upon Distributed Leadership theory, positing that leadership is a practice "stretched" across people and tools.

To operationalize these theories, we introduce two original mathematical models:

• Metric 1: The AI Access Burden Index (AABI) The AABI identifies the "stratification threshold" of AI access. It is calculated as:

  AABI = (Subscription Cost (C_sub) / Mean Monthly Disposable Income (I_disposable)) × 100%

  This ratio reveals the real-world economic pressure of AI adoption. While a $240 annual cost is negligible for high-income households, it represents a structural barrier for lower-income families, leading to a "Cognitive Tariff" that restricts access to high-reasoning models.

• Metric 2: The LEGA Decision Equation The LEGA framework evaluates institutional resilience through a multi-variable equation:

  LEGA = [Human Agency (ξ) × Institutional Governance Capacity (G_cap)] / [AABI × (1 + ln(1 + Cognitive Debt (△_gap)))]

  • Numerator (Drivers of Equity): Human Agency (ξ) represents the degree of effective intervention by educators in algorithmic workflows , while Governance Capacity (G_cap) denotes the institution's ability to implement mitigating strategies, such as district-level procurement or resource pooling.

  • Denominator (Barriers to Equity): AABI measures the economic hurdle. Cognitive Debt (△_gap) represents the cumulative disadvantage and "hallucination" risks incurred by students relegated to inferior, base-tier models.

  • Mathematical Logic: The use of a natural logarithm (ln) for Cognitive Debt reflects the non-linear, compounding nature of algorithmic disparity—where small initial gaps in the quality of AI assistance lead to exponentially widening achievement gaps over time.

3. European and International Dimension

This research provides a critical blueprint for global educational justice by synthesizing 2025/2026 data from the World Bank, OECD, and ITU. Our preliminary comparative modeling uncovers a profound injustice: the $20 monthly fee represents a negligible 0.4% of disposable income in affluent regions (e.g., USA), yet escalates to 13.3% in lower-income contexts (e.g., Nepal, Ethiopia). This 30-fold disparity signifies a new era of "Cognitive Colonialism". In alignment with the ECER 2026 theme of "Knowing and Acting," this study transforms the knowledge of the AABI into transformative actions through the Algorithmic Equity Toolkit, ensuring that academic success remains determined by merit and agency rather than household purchasing power.

Note to Reviewers: While I am submitting this as Ignite Talk to contribute to the high-level dialogue in NW 07, I am equally open to presenting this research in a Poster format should the committee find it more suitable for the program's interactive sessions.

Methodology, Research Instruments or Sources Used This study employs a Mixed-Methods Concurrent Triangulation design to integrate macro-socioeconomic modeling with micro-institutional analysis. 1. Quantitative Component: The AI Access Burden Index (AABI) The quantitative phase utilizes secondary data analysis from the 2024-2025 OECD "Education at a Glance" reports and World Bank Open Data. (1) The "$20 Barrier": The analysis is anchored to the $240 annual subscription cost ($20/month) of industry-standard premium models (e.g., ChatGPT Plus), representing the baseline for high-reasoning AI access. (2) AABI Calculation: The index is operationalized as: (Subscription Cost / Mean Monthly Disposable Income ) × 100%. This metric identifies the "stratification threshold" where AI access becomes a prohibitive "Cognitive Tariff". 2. Qualitative Component: Theoretical Calibration and Expert Consultation The researcher will participate in an academic week at Harvard University, Boston College, and NYU in April 2026 (Note: the original materials for this academic week are unrelated to the current research topic, which has been developed specifically for the ECER conference). During this period, the researcher will take the opportunity to engage in "Theoretical Stress-Testing" and informal consultations regarding the LEGA variables (Human Agency ξ, Governance Capacity Gcap, and Cognitive Debt Δgap) with international experts. This process involves seeking critical feedback on the model’s internal logic and variable weightings. Subsequently, the ECER conference will serve as the primary international forum to formally present and further refine the LEGA framework based on integrated scholarly insights. 3. Research Instrument: The LEGA Decision Equation The core instrument is the LEGA Framework, which evaluates institutional resilience via a non-linear equation: LEGA = [Agency (ξ) * Governance_Capacity (Gcap)] / [AABI * (1 + ln(1 + Cognitive_Debt (△gap)))]. Logic: A natural logarithm (ln) is applied to Cognitive Debt to simulate the non-linear, compounding nature of algorithmic disparity over time. 4. Data Synthesis and Simulation Quantitative data are processed via R/SPSS for correlation analysis. Qualitative insights are synthesized using NVivo. Finally, a Governance Simulation tests how changes in Gcap can mathematically offset the negative impact of AABI.

The research expects to deliver a 2026 Global Diagnosis of AI Access Burden. By applying the AABI Index, the study will provide the first empirical mapping of how the "$20 Intelligence Barrier" manifests as a "Cognitive Tariff" across diverse socioeconomic strata. This diagnostic report will serve as a baseline for identifying districts where algorithmic stratification has reached critical levels. A second primary outcome is the Validated LEGA Decision Model. The framework will first undergo a phase of theoretical stress-testing with experts at Harvard University, Boston College, and NYU in April 2026. Subsequently, the model will be presented at the ECER conference for a second round of high-level validation and peer-refinement through dialogues with the international expert community. This two-stage verification process ensures the model provides school administrators with a robust "health metric" to determine whether their current institutional resilience (Gcap) is sufficient to neutralize the algorithmic burdens (AABI) and cognitive debt (△gap) faced by their student populations. Furthermore, the study will culminate in the "Algorithmic Equity Toolkit" for school leaders. This practical manual translates mathematical variables—Human Agency (ξ), Governance Capacity (Gcap), and Cognitive Debt (△gap)—into actionable leadership strategies.Theoretically, this research extends the "Third-level Digital Divide" into the realm of algorithmic asset returns. By redefining the school leader as an "Institutional Architect," the project ensures that academic success remains determined by merit and agency (ξ) rather than household purchasing power.

Helsper, E. J. (2021). The Digital Disconnect: The Social Causes and Consequences of Digital Inequalities. SAGE Publications. Van Dijk, J. (2020). The Digital Divide. John Wiley & Sons. Van Deursen, A. J., & Van Dijk, J. A. (2015). The digital divide shifts to differences in usage. New Media & Society, 17(3), 376-391. Harris, A. (2013). Distributed Leadership in Practice. Springer Science & Business Media. Spillane, J. P. (2006). Distributed Leadership. Jossey-Bass. Rawls, J. (1999). A Theory of Justice (Revised ed.). Harvard University Press. Selwyn, N. (2024). Artificial Intelligence and the Future of Education: Critical Perspectives. Routledge. OECD. (2024). Education at a Glance 2024: OECD Indicators. OECD Publishing. UNESCO. (2023). Guidance for generative AI in education and research. UNESCO Publishing. UNESCO. (2025a). AI and the future of education: Disruptions, dilemmas and directions. UNESCO Publishing. UNESCO. (2025b). AI and education: Protecting the rights of learners. UNESCO Publishing. UNESCO. (2024). AI competency framework for teachers. UNESCO Publishing. OECD. (2025). Education at a Glance 2025: OECD Indicators. OECD Publishing. OECD. (2024). Education at a Glance 2024: OECD Indicators. OECD Publishing.