EducationSystematic Review

UDL Meets AI: Can Technology Finally Make Higher Education Truly Inclusive?

Universal Design for Learning provides the principles; generative AI provides the tools. Together, they could make higher education genuinely accessible to all learners. Four papers examine the promise and the implementation gap between UDL theory and classroom practice.

By Sean K.S. Shin
This blog summarizes research trends based on published paper abstracts. Specific numbers or findings may contain inaccuracies. For scholarly rigor, always consult the original papers cited in each post.

Universal Design for Learning rests on three principles: provide multiple means of engagement (the "why" of learning), multiple means of representation (the "what"), and multiple means of action and expression (the "how"). These principles, grounded in neuroscience research on learning variability, aim to create educational environments that work for all learnersβ€”not just those who fit the "average" student profile that most instruction implicitly assumes.

Generative AI offers new tools for implementing these principles at scale. AI can generate alternative representations of content (text to audio, text to visual, complex to simplified). It can provide personalized engagement pathways based on learner interests and goals. And it can accept diverse forms of student expression (voice, text, video, code) and evaluate them against learning outcomes rather than format compliance.

The question is whether institutions will use AI to implement UDL principles genuinelyβ€”or whether they will use AI to create the appearance of inclusion without the substance.

UDL Implementation for Students with Disabilities

Wahyuni, Pantiwati, and Sunaryo (2025) examine how UDL can enhance inclusive education for students with disabilities in higher education. Inclusive higher education is essential for ensuring equitable access to learning. UDL provides a flexible instructional framework that accommodates diverse learning needs and promotes inclusion.

The systematic review employing PRISMA methodology finds that while UDL principles are widely endorsed, implementation strategies vary significantly in quality and scope. Effective UDL implementation requires: institutional commitment (policies that mandate accessibility, not merely encourage it), faculty development (training in UDL principles and their practical application), technological infrastructure (platforms and tools that support multiple modalities), and ongoing assessment (evaluating whether UDL design actually improves outcomes for students with disabilities).

UDL Beyond the Classroom

Gilleran Stephens, Antwi, and Linnane (2025) apply UDL to redesign an environmental education outreach programβ€”demonstrating that UDL principles extend beyond formal classroom instruction to informal and non-formal education contexts. UDL is applied as a framework for creating a more inclusive and impactful experience.

The application of UDL to outreach and public engagement is significant because it challenges the assumption that UDL is relevant only to formal education. Science festivals, community workshops, museum programs, and public lectures can all benefit from multiple means of engagement, representation, and expression. The paper provides practical guidance on how to redesign existing programs using UDL principles without requiring complete program overhauls.

Dell'Anna, Marsili, and Bevilacqua (2025) examine the intersection of faculty development and UDL in advancing inclusion. Universities play a key role in promoting equal opportunities and accessibility, and the paper integrates two research areas: faculty development (encompassing teaching, professional development, and organizational dimensions) and UDL.

The paper argues that UDL implementation fails without faculty developmentβ€”a finding that echoes the teacher professional development literature reviewed earlier. Faculty who are trained in UDL principles but not in their practical application default to familiar instructional methods. Faculty who receive one-time UDL workshops but no ongoing support abandon new practices when challenges arise.

Effective faculty development for UDL requires: sustained engagement (not one-off workshops), discipline-specific application (UDL in engineering looks different from UDL in humanities), peer learning communities (faculty learning from each other's implementation experiences), and institutional incentives (recognizing UDL adoption in promotion and tenure decisions).

AI as UDL Enabler

Mallary, Moore, and McClain (2025) explore the transformative potential of integrating UDL with generative AI in adult and continuing higher education. By aligning UDL's principles with GenAI's capabilities in content generation, personalization, and accessibility, the paper proposes a framework for technology-enhanced inclusive instruction.

The AI-UDL integration framework identifies specific capabilities:

  • Multiple means of representation: AI generates alternative content formats (text summaries of videos, visual representations of quantitative data, simplified language versions of complex texts) automatically.
  • Multiple means of engagement: AI personalizes learning pathways based on learner interests, prior knowledge, and motivational preferences.
  • Multiple means of expression: AI accepts and evaluates diverse student outputs (voice recordings, visual presentations, written essays) against common learning outcomes.
The framework is conceptually promising but raises implementation concerns: AI-generated alternative representations may not capture the nuance of the original, AI personalization may encode biases that limit rather than expand options for marginalized learners, and AI evaluation of diverse expressions requires the kind of contextual judgment that current AI systems struggle with.

Claims and Evidence

<
ClaimEvidenceVerdict
UDL improves outcomes for students with disabilitiesWahyuni et al. (2025): evidence supports but implementation quality variesβœ… Supported (conditional on implementation)
UDL principles apply beyond formal classroom instructionGilleran Stephens et al. (2025): successful application to outreach programsβœ… Supported
Faculty development is necessary for UDL implementationDell'Anna et al. (2025): UDL fails without sustained faculty supportβœ… Supported
AI can effectively implement UDL at scaleMallary et al. (2025): framework proposed; empirical validation limited⚠️ Uncertain

Implications

The convergence of UDL and AI creates an opportunity to make inclusive education the default rather than the exception. But realizing this opportunity requires investment in faculty development, institutional infrastructure, and ongoing evaluationβ€”not merely the adoption of AI tools that claim to support accessibility. Technology can implement UDL principles, but only if the principles guide the technology rather than the technology defining the principles.

References (5)

[1] Wahyuni, S., Pantiwati, Y., & Sunaryo, H. (2025). Strategizing UDL Implementation: Inclusive Education for Students with Disabilities. Al-Ishlah, 17(1), 6630.
[2] Gilleran Stephens, C., Antwi, S.H., & Linnane, S. (2025). UDL: A Framework for Re-design of an Environmental Education Outreach Program. Discover Education, 4, 660.
[3] Dell'Anna, S., Marsili, F., & Bevilacqua, A. (2025). Faculty Development and Universal Design for Learning: advancing inclusion in higher education. Form@re, 25(1), 380–397.
[4] Mallary, K.J., Moore, E.J., & McClain, A.L. (2025). Artificial Intelligence and Universal Design for Learning: Transforming Teaching and Learning in Adult and Continuing Higher Education. New Directions for Adult and Continuing Education, 70016.
Wahyuni, S., Pantiwati, Y., Sunaryo, H., In'am, A., & Bastian, A. (2025). Strategizing Universal Design for Learning (UDL) Implementation: Enhancing Inclusive Education for Students with Disabilities in Higher Education. AL-ISHLAH: Jurnal Pendidikan, 17(1).

Explore this topic deeper

Search 290M+ papers, detect research gaps, and find what hasn't been studied yet.

Click to remove unwanted keywords

Search 8 keywords β†’