Generative AI as an Accommodation for Students with Learning Disabilities
ChatGPT Deep Research Content
Generative AI tools (like ChatGPT) are increasingly being used to support students with learning disabilities by modifying how tests and assessments are delivered or completed. Since late 2022, educators and researchers have reported numerous examples in both K-12 and higher education where AI was leveraged as an academic accommodation. Below we outline concrete cases – in the United States and internationally – including how these tools were implemented in assessment contexts and what outcomes were observed.
K-12 Education Use Cases
In primary and secondary schools, teachers have begun using generative AI to adapt assessments for students who need accommodations. This can involve simplifying language, adjusting question formats, or otherwise reducing cognitive load so that tests are more accessible:
Simplifying Test Content: AI chatbots can rephrase complex questions or reading passages into simpler language on the fly. For example, ChatGPT can take a long, dense text and summarize it at a lower reading level, which helps students with reading difficulties better comprehend exam materials. This use mirrors traditional accommodations like having a human reader or simplified text version, but AI offers it instantly and on demand.
Modifying Multiple-Choice Questions: Teachers have used ChatGPT to reduce the number of answer options in multiple-choice tests as an accommodation. One educator-facing report notes that for a student who struggles with too many choices, you can prompt ChatGPT to rewrite a quiz by removing one incorrect option per question. This results in a 3-choice question instead of 4, aligning with an established accommodation of limiting answer options to lessen confusion.
Generating Study Guides and Practice Questions: As a form of assessment support, AI can create personalized study materials for students with learning disabilities. Teachers input the upcoming test content and ask the AI to generate practice questions or review worksheets tailored to the student. This helps students prepare for tests in a structured way, reinforcing key concepts through repetition. For instance, ChatGPT’s strength in “productive repetition” means it can provide many examples or rephrasings of a concept – useful for a student with dyslexia who may need extra practice to master material.
Multi-language and Other Supports: In diverse classrooms, generative AI can instantly translate or reword assessment instructions. This benefits English language learners and also students with certain learning disabilities who might grasp a question better if it’s phrased differently. AI’s translation and rephrasing capabilities have been noted as a way to make test content more accessible without altering its meaning. While translation is not a traditional disability accommodation per se, it exemplifies how AI can flexibly meet individual needs (e.g. presenting a math problem in the student’s first language or in plainer English).
These K-12 implementations are often driven by teachers experimenting with AI to enhance Universal Design for Learning. Early anecdotal outcomes are promising – educators report improved student engagement and comprehension when assessment materials are tailored by AI to the right level. Importantly, these AI-assisted modifications mirror accommodations that might otherwise require significant teacher time (such as manually creating a simplified test version). With AI, the process is faster, allowing more students to get the modifications they need. However, the use of AI in K-12 accommodations is still in early stages, typically at the classroom level rather than mandated by district policy. Teachers are cautioned to review AI outputs for accuracy and appropriateness before giving them to students.
Higher Education Use Cases
Colleges and universities have also seen students and staff harness generative AI as an accommodation in assessments. This ranges from students using AI tools to assist their own test-taking or assignment completion, to institutions formally debating and piloting AI-based accommodations for exams. Below are several notable instances with outcomes:
AI-Assisted Writing and Grammar in Exams: In some cases, students with writing-related learning disabilities (such as dyslexia or dysgraphia) have been allowed to use AI writing assistants during exams as a disability accommodation. For example, a 2023 article in Language Assessment Quarterly notes that assistive writing tools – which include generative AI-based suggestions for wording and grammar – “are allowed as accommodations for students with learning disabilities on a case-by-case basis.” This means a university testing center might permit a student to use a tool like Grammarly or an AI-based autocomplete to help with spelling and sentence structure while writing an essay exam. The rationale is similar to allowing a human scribe or spell-checker: it helps the student demonstrate knowledge without being unfairly penalized for mechanics. The practice is still debated among educators (questions of fairness and exam validity arise), but it has precedent. In fact, a recent debate among language testing experts centered on whether test-takers should have access to generative AI tools during assessments to level the playing field. While no consensus has been reached, these discussions show higher-ed institutions are actively considering AI as a sanctioned accommodation in test settings.
Case Study – Dyslexic Student’s Success with ChatGPT: A well-publicized example comes from France: Myriem Khal, a dyslexic college student, used ChatGPT as a study aid and saw dramatic improvements in her exam performance. Khal, who studies computer engineering, struggled to read and write in English (the language of instruction for some of her courses) due to dyslexia. She began using ChatGPT to “restructure course materials” and explain tough concepts in simpler terms. Essentially, she would input lecture notes or textbook passages and have the AI summarize or clarify them. The outcome was striking – Khal “passed her final exams with flying colors” after using the AI tool to support her studying. In interviews, she reported getting “great marks” on those exams, a success she directly attributes to ChatGPT helping her understand the material better. This case illustrates the potential of generative AI to function like a personalized tutor or text reader for a student with a learning disability. It also highlights that the implementation was student-driven: Khal used ChatGPT on her own initiative as an allowed study resource, not as a formally provided accommodation. Nonetheless, the positive outcome – improved exam scores and reduced frustration – underscores how AI can empower students to overcome disability-related challenges in assessments.
University Disability Services Adopting AI: Some disability support offices in higher ed are now integrating AI tools into their accommodation toolbox. In Ontario, Canada, campus accessibility staff reported that students with hearing or writing impairments have used AI transcription services (like Otter.ai) to transcribe lectures and students with reading-based learning disabilities have used ChatGPT’s speech-to-text and summarization features to assist with coursework. This means a student who has trouble taking notes due to dysgraphia can record a lecture and have an AI produce an accurate transcript, or a student with dyslexia can dictate an essay answer and have AI convert it to text. These implementations effectively serve as accommodations during assessments and study sessions – providing “quick access to accommodations” without always needing human intermediaries. An important reported outcome is that AI tools reduced the workload on disability services staff, who otherwise would be arranging note-takers, human transcribers, or creating alternate formats. By automating parts of the accommodation, students got support faster and staff could focus on cases requiring personal intervention. Universities are also exploring AI for accessible content creation – for instance, using generative AI as an “accessibility checker” to review exam materials, syllabi, or presentation slides and suggest adjustments (like adding alt-text, simplifying phrasing, or formatting for screen readers). This proactive use of AI aims to ensure assessments are designed to be inclusive from the start, benefiting all students.
Autism and ADHD Support in Assignments: Beyond formal exams, generative AI has been used by students with neurodiversity to navigate academic assignments, which are a form of assessment. A 2024 research study conducted interviews with 33 students with disabilities across Europe to see how they use ChatGPT in higher education. The results showed many “significant opportunities [for ChatGPT] as an assistant in teaching, writing, reading and research” for these students. For example, one autistic university student described how they used ChatGPT to summarize an overwhelming reading assignment into the core points, enabling them to grasp the material in far less time. Others used ChatGPT to help outline essays or suggest opening sentences, which is akin to an organizational accommodation for those with ADHD or other learning disabilities who struggle to initiate writing. These students reported that AI assistance improved their understanding and reduced the effort needed to complete coursework. Notably, one participant with ADHD said ChatGPT was extremely useful for breaking through “the terror of the blank page” by suggesting ways to start writing, thereby helping them overcome executive functioning barriers when working on an assessed essay. Although these uses were not always through official disability accommodations (sometimes they were self-help strategies), they directly supported the students in assessment situations (e.g. writing papers, understanding exam readings) and led to better performance and confidence. Researchers recommend that universities recognize these benefits and consider them when crafting AI policies – ensuring that students who need such tools can use them without penalty.
Policy and Trials: In the U.S., some colleges are cautiously piloting the idea of AI in assessments for equity reasons. A few professors have run controlled trials – for instance, allowing an AI tool for one portion of an exam or assignment – to observe the effects. While results of such trials are still emerging, they feed into broader policy discussions. The National Center for Learning Disabilities (NCLD) has noted that disability advocates see value in AI tools “leveling the playing field” if used ethically. They argue that a one-size-fits-all ban on AI may ignore the legitimate needs of students who require extra support. This has led to some universities issuing nuanced guidelines, where AI can be used by students with documented accommodations (similar to how a calculator or spell-check might be allowed for those who need it, even if it’s barred for others). The key is ensuring fairness: if AI is an accommodation, it would typically require disability services approval to prevent abuse by those without disabilities. These early policy responses in higher ed are attempting to balance academic integrity with accessibility. We are seeing the beginnings of formal recognition that generative AI, when used under supervision, “could be an accommodation on its own” for certain tasks – though some worry this might be unfair if only some students are authorized to use it. The conversation is ongoing, with outcomes from pilot implementations informing future rules.
Outcomes and Considerations
The concrete outcomes from using generative AI as an accommodation have been encouraging but mixed. On the positive side, many of the above cases report improved academic performance, increased independence, and time savings. Students like Myriem Khal achieved higher exam scores once AI helped make course content understandable. Interviewed students with disabilities said AI tools gave them more confidence and reduced stress when completing assessments, since they could get assistance without feeling judged. Disability services staff also observed faster service delivery (e.g. immediate lecture transcripts) and possibly a reduction in last-minute accommodation requests when students can self-serve with AI. In short, when implemented thoughtfully, generative AI has helped level the academic playing field by mitigating some disability-related barriers in testing and assignments.
However, there are important caveats and concerns noted in the sources. Educators warn that AI is not a panacea: for instance, learning disability experts worry about over-reliance on ChatGPT as a “catch-all solution” without teaching underlying skills. There are also equity questions – if only students with formal diagnoses can use AI on an exam, is that fair to other students, or is it a necessary distinction? Some disabled students themselves voiced that AI’s role in formal assessments should be limited to preserve academic integrity. In a UK survey, both disabled and non-disabled students generally felt AI should be used sparingly or not at all in graded assessments, even though they welcomed it for studying and support. This indicates a desire to avoid any hint of unfair advantage or dependency during exams. Conversely, disability advocates caution that in the rush to “AI-proof” exams (for example, banning take-home essays in favor of in-person written tests), institutions might inadvertently make assessments less inclusive. A rigid anti-AI stance can collide with the need for accommodations – for example, a student with dyslexia might genuinely need a tool to read or compose text, yet a blanket ban would forbid it. Kevin Boyle, an accessibility professional, noted that when instructors react to ChatGPT by imposing strict in-class writing conditions, “students who require accommodations are already working uphill… overreacting [to AI] is increasing the slope of those hills.”. In other words, poorly calibrated policies could hurt the very students accommodations are meant to help.
Going forward, the emerging consensus is that context and oversight matter. Generative AI can absolutely support students with learning disabilities in tests and assessments, as shown by the cases above, but it should be used in alignment with individualized needs and academic guidelines. Many colleges are drafting policies to explicitly permit AI usage as an accommodation with prior approval. K-12 schools, meanwhile, are developing teacher training on using AI for accessibility, while emphasizing that any AI-modified test still needs human review for quality and alignment with learning goals. There is also a call for AI developers to build in disability features (e.g. “accessible mode” for an educational AI app) and for involving users with disabilities in design.
In summary, post-2022 experiences show generative AI can be a powerful tool to modify and support academic assessments for learners with disabilities. Concrete implementations range from auto-simplifying exam questions and providing on-demand reading assistance, to permitting AI-based grammar support during exams or using chatbots to generate practice quizzes for students with ADHD. The outcomes – such as higher test scores, better engagement with material, and faster access to accommodations – are highly encouraging. At the same time, educators are proceeding carefully, ensuring that the use of AI remains fair and transparent. Both in the U.S. and internationally, institutions are pilot-testing these accommodations and sharing best practices. As one education expert put it, outright AI bans are “a one-size-fits-all approach [that] may fail to create inclusive and supportive academic environments”. Instead, when used ethically, generative AI might be the next step in providing personalized accommodations – ultimately helping students with learning disabilities demonstrate their true abilities in assessments without the barriers that once held them back.
Sources:
Ayala, S. (2023). ChatGPT as a Universal Design for Learning Tool – Supporting College Students with Disabilities. Educational Renaissance, 12. (Case examples of ChatGPT aiding a student with autism in research and writing)
Voss, E. et al. (2023). The Use of Assistive Technologies Including Generative AI by Test Takers in Language Assessment: A Debate of Theory and Practice. Language Assessment Quarterly, 20(4-5), 520-532. (Notes that AI writing tools can be allowed as accommodations for learning-disabled students in exams)
NGPF (2023). How ChatGPT Revolutionizes Teaching for Diverse Learners. (K-12 teacher resource describing ChatGPT prompts to adapt tests, e.g. fewer multiple-choice options for accommodations)
Chronicle of Higher Education (Feb 2023). How ChatGPT Could Help or Hurt Students With Disabilities. (Insights from disability support experts on using AI for note-taking, social scripting, study help; cautions against blanket bans)
Rocky Mountain ADA Center (Bezyak, J. et al., 2023). ChatGPT and Disability: Benefits, Concerns, and Future Potential. (Highlights that AI tools like ChatGPT can increase reading comprehension, assist in writing, and potentially “level the playing field” for students with disabilities)
Business Insider (Mar 2023). “A student is using ChatGPT to overcome dyslexia… passed her exams with flying colors.” (Story of Myriem Khal, French student with dyslexia, who used ChatGPT to restructure course content and achieved top exam results)
Nottingham Trent University (Jul 2024). Research shows risks and opportunities of GenAI for dyslexic, neurodivergent and disabled students. (Survey finding 59% of disabled students had used GenAI to explain concepts; discusses student attitudes toward AI in assessments)
Higher Education Quality Council of Ontario – HEQCO (Nov 2024). “Using Generative AI to Make Learning More Accessible.” (Report on Canadian post-secondary students; notes students using AI transcription and summarization as accommodations, and instructors allowing AI in limited assignment contexts)
Pierrès, O. et al. (2024). Exploring the Role of Generative AI in Higher Education: Interviews with Students with Disabilities. Education and Information Technologies, 30(2025). (European study; documents students using ChatGPT to summarize readings, plan essays, and manage study schedules, with beneficial effects on their learning outcomes)