Student or educator? Sign up with your school email and get 6 months of Pro free.
Cited by 4Artificial Intelligence in EducationLegal Issues in EducationDisability Education and Employment

The DOI matches its Crossref record; 28 of the 28 references it cites were found in a registry or an index; none of the 73 sources its claims cite is checked for support yet.

Reasonable adjustments for neurodivergent students in higher education: generative AI, accessibility, and mediated authorship

Open access
Morgan, Daniel·2026·Disability & Society

Abstract

Neurodivergent students face significant disadvantages in higher education, where assessment often privileges neuro-normative written expression. This article argues that Large Language Models (LLMs) should be framed as accessibility tools that scaffold expression while preserving the student’s intellectual authorship, reframing the debate from academic integrity to equity. The argument is grounded in UK equality law and contends that the Court of Appeal’s decision in University of Bristol v Abrahart requires universities to distinguish competence standards from the methods used to assess them. On this basis, the regulated use of LLMs may amount to a reasonable adjustment where a student’s barrier lies in written expression rather than reasoning. The article advances a four-pillar model for responsible integration, addressing risks such as algorithmic bias, opacity, and data privacy. It concludes that LLMs, situated within traditions of mediated authorship, can support inclusion without displacing academic standards.Points of interestThis article highlights how university assessments often focus on perfect writing, which can create unfair barriers for students with dyslexia, autism, or ADHD.It argues that Artificial Intelligence (AI) writing tools, including Large Language Models (LLMs), should be viewed as accessibility aids that help students organise their thoughts and express their ideas clearly.It introduces the idea of ‘mediated authorship’ to explain that AI support can be compatible with academic integrity. The student provides the ideas and arguments, while the tool helps improve clarity and structure.It argues that, under UK equality law, permitting regulated use of AI tools may be a ‘reasonable adjustment’ for some disabled students.It concludes that this approach helps ensure students are assessed on their knowledge and critical thinking, rather than on their ability to match a narrow style of academic writing.

Overview

CoChat Summary

TL;DR

This is a conceptual socio-legal argument that Large Language Models (LLMs) can function as accessibility tools for neurodivergent students in higher education, rather than primarily as threats to academic integrity. The paper claims that, under UK equality law, regulated LLM use may count as a reasonable adjustment when the barrier is written expression rather than reasoning.

Key findings

  • The paper argues that assessment in higher education often privileges “linear, polished” writing and thereby creates structural disadvantage for autistic, dyslexic, and ADHD students; it states that this is not a deficit of thought but a mismatch between cognition and institutional demand.
  • It cites the Court of Appeal decision in University of Bristol v Abrahart [2024] as strengthening the anticipatory duty to make reasonable adjustments, and says a blanket prohibition on LLMs is “legally precarious” when a student’s neurodivergence creates a substantial barrier to conventional prose.
  • It proposes a four-pillar model for responsible LLM integration: Institutional Provision, Transparency Frameworks, Assessment Redesign, and Training and Safeguards.
  • The paper reports specific empirical cautions: Goodman et al. (2022) had a small sample size, experienced participant dropout, and focused on email writing rather than academic assessment; Goodman et al. (2024) concluded current LLMs may fall short of the accuracy and quality thresholds required for high-stakes writing.

Methods

  • This is a socio-legal and conceptual analysis, not an empirical study.
  • It synthesises existing literature on neurodivergent students’ writing barriers, assistive technologies, and emerging LLM use in higher education, while grounding the argument in the Equality Act 2010, the Abrahart judgment, and the UNCRPD.
  • The main analysis is normative and interpretive: it reframes LLMs through the concept of mediated authorship and evaluates legal, pedagogical, and accessibility implications.

Limitations

  • The paper itself acknowledges that the empirical evidence on LLMs for accessibility is preliminary and that findings from prior studies should be treated with caution because of limits such as small samples, dropout, and tasks that may not match academic assessment.
  • It notes that LLMs introduce risks of bias, opacity, misinformation, dependency, and data privacy concerns, and says these cannot be eliminated, only managed.
  • Evident from the text: because the article is a conceptual argument rather than a primary study, it does not provide original participant data or direct evaluation of the proposed four-pillar model.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1080/09687599.2026.2667528

Cites

28 of 28 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

0 Valid · 15 Not valid · 62 Not checked · 83 Own results and interpretations

48 more claims

Create a free account to see all 160 claims and every source verdict.

See all claims — free account
  • Probably not a substantial claim. 90
  • It may state two separate claims as one. 41

Create a free account to see these claims.

Citations

Cites

Show all 27 references

Research works printed but not linked: 5

  1. Bathaee, Y. 2018. “The Artificial Intelligence Black Box and the Failure of Intent and Causation.” Harvard Journal of Law and Technology 31 (2): 890–938.
  2. Nussbaum, M. C. 2007. “Human Rights and Human Capabilities.” Harvard Human Rights Journal 20: 21–24.
  3. Singer, J. 1999. “‘Why Can’t You be Normal for Once in Your Life?’ From a ‘Problem with no Name’ to the Emergence of a New Category of Difference.” In Disability Discourse, edited by M. Corker and S. French, 59–67. Buckingham: Open University Press.
  4. United Nations. 2006. “Convention on the Rights of Persons with Disabilities (CRPD).” Treaty Series 2515: 3.
  5. Walker, N. 2021. Neuroqueer Heresies: Notes on the Neurodiversity Paradigm, Autistic Empowerment, and Postnormal Possibilities. Fort Worth: Autonomous Press.

Metadata

Authors
Morgan, D.
Affiliations
University of Wolverhampton (GB)
Corresponding author
Morgan, Daniel
Venue
Disability & Society
Published
2026-05-11
Volume
41
Issue
7
Pages
1836-1856
ISSN
  • 0267-4645 (ISSN-L)
  • 0968-7599
  • 1360-0508
OpenAlex
W7160840067
Cited by
4
References
27
Citations by year
2026: 4
FWCI
34.03
Type
journal article
Language
en
Open access
Licence
Locations
  • Disability & Society (publishedVersion, cc-by-nc-nd)
  • SSRN Electronic Journal (acceptedVersion)
  • Wolverhampton Intellectual Repository and E-Theses (University of Wolverhampton) (submittedVersion)
Topics
  • Social Sciences › Education › Artificial Intelligence in Education
  • Social Sciences › Law › Legal Issues in Education
  • Social Sciences › Education › Disability Education and Employment
Keywords
neurodivergent students, reasonable accommodation, large language models, generative AI in education, accessibility, academic integrity, dyslexia, autism spectrum disorder, attention deficit hyperactivity disorder, higher education assessment, algorithmic bias, data privacy, academic writing

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

A language model drew the claims from the paper on 2026-10-06, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-06 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Reasonable adjustments for neurodivergent students in higher education: generative AI, accessibility, and mediated authorship. https://app.cochat.ai/papers/10.1080/09687599.2026.2667528
The paper
Morgan, D. (2026). Reasonable adjustments for neurodivergent students in higher education: generative AI, accessibility, and mediated authorship. Disability & Society. https://doi.org/10.1080/09687599.2026.2667528

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.