AI in special education is growing fast, and it is growing ahead of the rules. In 2022–23, 7.5 million students, 15 percent of all public school students, received special education services under the Individuals with Disabilities Education Act (IDEA), and about 5.3 million were English learners. AI tools can help many of these diverse learners by reading text aloud, simplifying language, and adjusting practice to each student. Teachers are also using AI on the plans that govern students’ services: in 2024–25, 57 percent of teachers said they had used AI to help develop an IEP or 504 plan, up from 39 percent the year before, according to the Center for Democracy & Technology.
This article explains how AI can support students with disabilities and other diverse learners, where it raises risks, and what schools should put in place.
How AI can support diverse learners
Accessibility and assistive technology
Many of the most useful tools build on long-standing assistive technology:
- Reading support. Microsoft Immersive Reader, free in Microsoft 365 apps, reads text aloud while highlighting each word, breaks words into syllables and adjusts spacing and line focus.
- Text-to-speech and simplification. Read&Write combines text-to-speech, speech-to-text, word prediction and translation, and uses AI to simplify text and produce natural-sounding voices. Free premium accounts are offered to K–12 teachers.
- Captions and transcripts. Captions improve comprehension and memory, especially for students who are deaf or hard of hearing and students learning in a second language.
Personalized practice
Adaptive programs adjust difficulty and pacing based on each student’s responses, which can help students who need more practice, and students ready to move faster, in the same class.
Reducing the load
CAST, the organization behind Universal Design for Learning (UDL), describes AI as potentially working like “assistive intelligence”: lowering the cognitive load of a task, for example by changing a text’s reading level, so students can focus on learning. CAST urges that AI tools be designed with UDL from the start and that educators stay “in the loop.” In a CAST survey, 70 percent of educators said they wanted to learn how to use AI to support students with disabilities and English learners, and nearly two-thirds said they were not getting the help they need.
AI and IEPs: use with care
An individualized education program (IEP) is a legal plan built around one student. The Center for Democracy & Technology found that in 2024–25, 15 percent of teachers used AI to write an IEP or 504 plan in full, up from 8 percent the year before. It recommends that teachers not enter students’ personal information into an AI tool unless the school has an agreement with the vendor, and that schools require human review of anything AI produces.
Common Sense Media goes further. After testing AI teacher assistants in 2025, it rated them moderate risk, warned of bias based on perceived student backgrounds in IEP and behavior-plan suggestions, and advised against using AI to write IEPs or behavior plans at all.
Risks for students with disabilities
- Privacy. 60 percent of students with an IEP or 504 plan said they were concerned about student privacy, compared with 42 percent of other students, in a 2025 CDT survey.
- Cheating accusations and discipline. In 2023–24, 64 percent of teachers said students had gotten in trouble for using, or being accused of using, generative AI, and licensed special education teachers were especially likely to use AI detection tools, CDT reported. It warned that students with IEPs and 504 plans, who use AI more, may face greater risk.
- Biased or generic outputs. AI suggestions can reflect assumptions about students rather than their actual needs, which is why human review matters.
What federal guidance says
In July 2025, the U.S. Department of Education said in guidance on using federal funds for AI that AI tools should be accessible to students, educators and families with disabilities. Schools remain responsible under IDEA and Section 504 for the services each student receives, whatever tools are used to plan them.
What schools should put in place
- Keep people responsible for plans. IEP teams, not software, decide services and goals. Any AI draft should be checked against the student’s actual evaluations and needs.
- Protect student data. Use only tools covered by a district privacy agreement, and never paste identifiable student information into a public chatbot.
- Use AI for access first. Read-aloud, captions, translation and text simplification are lower-risk uses with clear benefits.
- Be careful with detectors. Do not discipline students on the basis of an AI detector alone.
- Train staff. Teachers want help using AI for students with disabilities; give it to them.
Used carefully, AI can make classrooms more accessible for students with disabilities and English learners. Used carelessly, it can make plans less individual and put students’ privacy at risk.
Related reading
Subscribe to edCircuit to stay up to date on all of our shows, podcasts, news, and thought leadership articles.
More on Special Education and IEPs
