Reimagining Accountability: Using ChatGPT to Empower Parents/Stakeholders and Advance the Unfinished Equity Agenda of No Child Left Behind

This study examined how generative artificial intelligence (AI), specifically ChatGPT, functioned as an interpretive intermediary to support stakeholder engagement with complex educational accountability data. Guided by Bronfenbrenner’s ecological systems theory and principles of educational data transparency, the study analyzed district-level thirdgrade mathematics performance data from 32 New York City community school districts representing 51,210 third-grade students. A descriptive and correlational design was employed to examine patterns of advanced achievement (Level 4) and underperformance (Level 1) across districts by socioeconomic status (SES), race/ethnicity, and gender. Statistical analyses included descriptive statistics, Pearson correlations, and comparative district-level analyses conducted and verified using Microsoft Excel 365 Data Analysis ToolPak. Findings revealed substantial variation in mathematics outcomes across districts. Underperformance was concentrated in districts serving predominantly Hispanic/Latinx and Black student populations, while district SES was strongly associated with both advanced achievement and underperformance. Latinx students demonstrated lower Level 4 achievement rates than Asian and White peers, even in some relatively advantaged districts. Boys outperformed girls in advanced mathematics achievement in 21 of 25 districts, suggesting an early gender gap. Generative AI translated statistical outputs into accessible explanations and facilitated structured interpretation using a 5Ws framework. All AI-assisted outputs were validated against the original dataset. The findings suggest generative AI can enhance accessibility, stakeholder sensemaking, and engagement with educational accountability data.