Code for Equity: A Human-AI Process for Reviewing Education Policy Against CARE Frameworks

This article presents a methodology combining AI-powered text mining with culturally responsive, antiracist, and equitable (CARE) principles to analyze educational policy. Using sentiment and word-count analyses, we evaluate policies from four U.S. states (North Carolina, California, Oregon, and Washington) and federal legislation. Findings reveal that procedural, compliance-focused language predominates, limiting transformative potential—patterns mirrored in school and district implementation challenges. California demonstrates stronger culturally responsive integration. We offer a replicable rubric that researchers, educators, and policymakers can use to evaluate policy design and identify implementation barriers. This approach provides an accessible method for equity-centered policy review while addressing ethical considerations inherent in AI-assisted analysis, including transparency about tool limitations and the primacy of researcher judgment over automated outputs. This methodology extends a prior qualitative policy analysis (Ouedraogo-Thomas et al., 2025) into a reproducible computational framework applicable across policy contexts.