Source: Dusseault, B., Sims, M., & Berardino, M. (2025). AI early adopter districts: The promises and challenges of using AI to transform education. Center on Reinventing Public Education. https://files.eric.ed.gov/fulltext/ED674608.pdf
LEARN Brief Credit: Dr. Jeannie Haubert
Overview
This is brief 4 of 4 in a comprehensive national study of AI adoption at the district level. CRPE’s national study of AI Early Adopter districts closes with both a warning and a plan. AI tools today are rarely designed with multilingual learners or students with disabilities in mind, and districts risk leaving these students further behind unless equity is built into AI strategy from the very start — not bolted on after the fact. This brief translates the study’s findings on special populations, plus its three headline recommendations, into a concrete action plan superintendents can put in place this school year.
Key Insights
- Few AI tools are built with multilingual learners (MLLs) or students with disabilities in mind; most current applications are narrow — translation or simplified text generation — rather than holistic support woven into everyday instruction and IEP or 504 planning.
- Translation tools are a genuine bright spot: educators report that AI-powered translation has helped humanize relationships with MLL families, opened up communication in schools serving dozens of languages, and in some cases enabled more rigorous assignments as teachers better understood what students were actually capable of.
- Districts consistently state equity goals but rarely audit whether their AI tools actually help or quietly widen access gaps, and understandable caution about testing unproven tools on the students with the most specific needs can itself become a barrier — leaving these learners under-served twice over.
- Some early models point to a better path: one district learned through piloting that an AI tutoring tool worked especially well for one specific learner profile, and had teachers “co-teach” alongside the AI tool for those students — freeing teachers to give more personalized, human attention to higher-need students who needed it most.
- CRPE’s three headline recommendations for the field: (1) design AI strategies to solve systemic instructional and equity problems — tools alone can’t fix broken schedules, misaligned curricula, or disconnected data systems; (2) replace fragmented, ad hoc policies with a coherent, cross-functional strategy and clear guardrails; and (3) build evidence faster through networked training, shared research, and rapid-cycle learning across districts rather than each one testing tools in isolation.
Action Steps
- Require every new AI pilot proposal to explicitly name how it will serve multilingual learners and students with disabilities — not just general education students — before it’s approved.
- Stand up a cross-functional AI governance team that includes technology, curriculum, special education, and multilingual services leaders at the same table from day one.
- Audit your current AI tools specifically for equity impact: are they closing access gaps, or automating around the students who already receive the fewest resources?
- Join or help start a peer network to share real-time evidence on what’s working for underserved learners, rather than waiting years for conclusive research.
- Set clear, written guardrails for responsible AI use with special populations, then give staff explicit permission to pilot within them — caution and progress aren’t mutually exclusive.
Full Article
Read the full CRPE report: https://files.eric.ed.gov/fulltext/ED674608.pdf
