NPR: special ed teachers use AI

NPR reported on May 20 that special-education teachers in the United States are increasingly using artificial intelligence to draft individualized education programs, parent communications and other required paperwork, as schools struggle with shortages and burnout. The story centered on Mary Acebu, a special-education teacher at Riverview Middle School in Bay Point, California, who told NPR that AI has helped her spend less time at a computer and more time with students. The shift described in the report was narrow: teachers are using AI as a first-draft tool, not to make final decisions about services or progress. At the same time, a separate dispute at Western University in Ontario has added to pressure for more in-person, observable assessment after a professor alleged widespread AI cheating on a final exam. ### Why are special-education teachers reaching for AI now? The paperwork load in special education has been a central reason teachers cite for stress and attrition. NPR reported that Acebu used to arrive at school by 6:30 a.m. and often leave after dark with paperwork still unfinished. She said that changed after two years of experimenting with AI tools to move through documentation faster. (tspr.org) The shortage data helps explain the appeal. NPR said 45 states reported special-education teacher shortages in the 2024-25 school year, and turnover was worse in schools serving largely low-income students. A Learning Policy Institute analysis of federal data also found special education was the most commonly reported shortage area, with 45 states identifying it. (tspr.org) ### What exactly are teachers using AI to do? IEPs are the main task in view. NPR said special educators are using AI to speed up drafting of individualized education programs, which set out goals, services and supports for students with disabilities. The report described AI as a way to generate or customize language that teachers then review and revise. (tspr.org) The scale of use has risen quickly. The Center for Democracy and Technology said 57% of teachers reported using AI to develop an IEP or 504 plan during the 2024-2025 school year, up 18 percentage points from the previous year’s 39%. CDT also said the practice raises concerns tied to student privacy law and disability law, including FERPA and IDEA compliance. (tspr.org) ### Are researchers saying the output is good enough to trust? University researchers cited by NPR said AI-assisted drafting can, when used appropriately, match or exceed the quality of documents teachers produce on their own. NPR attributed that finding to work involving the University of Virginia and the University of Central Florida. (cdt.org) Olivia Coleman of the University of Central Florida told NPR that more direct time between a student with a disability and a teacher “often yields better outcomes,” linking the administrative time savings to student support rather than automation for its own sake. That framing matters because the teachers in NPR’s report were still described as the people making the final judgments. (tspr.org) ### Where are the limits and risks? CDT said the legal and ethical risks are substantial if schools feed sensitive student information into outside systems without proper safeguards. Its briefing flagged privacy, disability-rights and compliance issues, which means the convenience of faster drafting does not remove the need for district policy, review and human signoff. (tspr.org) NPR’s account also did not describe AI as replacing teacher expertise. The teachers in the story used it to produce drafts and organize language, while the responsibility for deciding goals, services and progress remained with educators and families. ### Why does the Western University case matter to this conversation? CTV News reported on May 20 that a Western University professor alleged most students in one of his classes used AI to cheat on a final exam and said he rejected the exam results. (cdt.org) Western University said it was reviewing the matter. The report points to a separate but related issue in education: if generative AI makes take-home or typed work harder to verify, schools may put more weight on live, in-class demonstrations of knowledge. (tspr.org) The contrast is straightforward. In the NPR report, AI appeared on the teacher side as an administrative aid under human review. In the Western case, AI appeared on the student side as a suspected integrity problem in assessment. Those two developments together are likely to keep schools focused on one distinction: automation for drafting is being tested, but judgment and verification are staying with humans. (youtube.com) Western University said it was reviewing the allegations reported by CTV, while the NPR story pointed to continued teacher use of AI in the 2025-26 school year as districts and researchers work through privacy, policy and quality questions. (tspr.org)