Designing AI for All: Why Disability Must Lead the Conversation
By Aurora Dreger, Educating All Learners Alliance
How do we ensure students with disabilities are considered from the beginning, not as an add-on or afterthought? This question is the catalyst for all our work within the Educating All Learners Alliance (EALA). In fact, this axiom is the driving force for our recent work in AI policy. In April 2026, EALA and New America released the joint policy brief Prioritizing Students with Disabilities in AI Policy, which raised an even more important question: “Who gets left behind if we get this wrong?”
Caption: Students and parents participating in an AI co-design session at their local school.
AI is now a reality in education. As of the 2024-2025 school year, 57% of licensed special education teachers reported using AI to help develop IEPs or 504 plans, according to a recent study from the Center for Democracy and Technology. AI can offer more personalized, accessible learning for the 1 in 5 students with learning disabilities. However, without strong oversight, AI can also misidentify needs, lower standards, and embed bias in ways that are hard to detect and challenge. These risks are urgent for policymakers, educators, students, and families.
EALA refuses to treat disability as an afterthought. Students with disabilities and learning differences must be central in designing and governing AI systems. AI literacy is essential for safe, responsible use.We discovered additional insights from our work as we continue to prioritize these students:
Design with students, not just for students: Students with disabilities, along with their families and educators, must be involved in developing, testing, and evaluating AI tools. This is not only about representation; it is about building systems that actually work in real-world contexts.
Mandate transparency and accountability: AI systems in education should be clear and open to review. When an algorithm recommends placement, support, or performance, educators and families need to know how and why it came to that decision. This is more than human oversight when it comes to decision-making; it is also about the technology’s built-in transparency and audit capability.
Prioritize accessibility as a baseline, not a feature: Accessibility cannot be optional. Systems must support diverse learning needs, including assistive technologies and adherence to Universal Design for Learning principles, from the start.
Protect student data with elevated care: Students with disabilities often have more detailed and sensitive data, including health information. Our AI Policy brief calls for stronger protections to prevent misuse. This is especially important as AI systems depend heavily on data collection and could link complex student data profiles to consequential decision-making.
Invest in educator capacity: Even the best-designed tools won’t work without proper use. Educators need training on how to use and evaluate AI tools as well as determine which tools are best in an educational context. This includes the ability to identify bias, know the tool's limits, and advocate for their students.
It is clear that inclusive AI policy is not a niche concern. It is a litmus test for the integrity of our entire education system. If AI serves students with disabilities well, it will likely work for all students; whereas if the system overlooks these students, it is fundamentally flawed.
EALA unites diverse voices to find real solutions for students with learning differences. Inclusive collaboration shaped the Prioritizing Students with Disabilities in AI Policy paper and drives our support for National AI Literacy Day. As AI transforms education, EALA accepts the challenge to expand opportunity—not deepen inequality.
