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 learnersβnot 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 genuinelyβor 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 programβdemonstrating 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.
Faculty Development: The Missing Link
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 developmentβa 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
<
| Claim | Evidence | Verdict |
|---|
| UDL improves outcomes for students with disabilities | Wahyuni et al. (2025): evidence supports but implementation quality varies | β
Supported (conditional on implementation) |
| UDL principles apply beyond formal classroom instruction | Gilleran Stephens et al. (2025): successful application to outreach programs | β
Supported |
| Faculty development is necessary for UDL implementation | Dell'Anna et al. (2025): UDL fails without sustained faculty support | β
Supported |
| AI can effectively implement UDL at scale | Mallary et al. (2025): framework proposed; empirical validation limited | β οΈ 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 evaluationβnot 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.
λ©΄μ±
μ‘°ν: μ΄ κ²μλ¬Όμ μ 보 μ 곡μ λͺ©μ μΌλ‘ ν μ°κ΅¬ λν₯ κ°μμ΄λ€. νμ μ°κ΅¬μμ μΈμ©νκΈ° μ μ ꡬ체μ μΈ μ°κ΅¬ κ²°κ³Ό, ν΅κ³ λ° μ£Όμ₯μ μλ³Έ λ
Όλ¬Έμ ν΅ν΄ κ²μ¦ν΄μΌ νλ€.
UDLκ³Ό AIμ λ§λ¨: κΈ°μ μ λ§μΉ¨λ΄ κ³ λ±κ΅μ‘μ μ§μ μΌλ‘ ν¬μ©μ μΌλ‘ λ§λ€ μ μλκ°?
보νΈμ νμ΅ μ€κ³(Universal Design for Learning, UDL)λ μΈ κ°μ§ μμΉμ κΈ°λ°νλ€: λ€μν μ°Έμ¬ μλ¨ μ 곡(νμ΅μ "μ΄μ "), λ€μν νν μλ¨ μ 곡(νμ΅μ "무μ"), κ·Έλ¦¬κ³ λ€μν νλ λ° νν μλ¨ μ 곡(νμ΅μ "λ°©λ²"). νμ΅ λ€μμ±μ κ΄ν μ κ²½κ³Όν μ°κ΅¬μ κ·Όκ±°ν μ΄λ¬ν μμΉλ€μ, λλΆλΆμ κ΅μλ²μ΄ μ묡μ μΌλ‘ μ μ νλ "νκ· μ μΈ" νμ μ νμ λ§λ νμ΅μλ§μ΄ μλλΌ λͺ¨λ νμ΅μμκ² ν¨κ³Όμ μΈ κ΅μ‘ νκ²½μ μ‘°μ±νλ κ²μ λͺ©νλ‘ νλ€.
μμ±ν AIλ μ΄λ¬ν μμΉλ€μ λκ·λͺ¨λ‘ ꡬννκΈ° μν μλ‘μ΄ λꡬλ₯Ό μ 곡νλ€. AIλ μ½ν
μΈ μ λ체 νν λ°©μμ μμ±ν μ μλ€(ν
μ€νΈλ₯Ό μ€λμ€λ‘, ν
μ€νΈλ₯Ό μκ° μλ£λ‘, 볡μ‘ν λ΄μ©μ λ¨μνλ ννλ‘). λν νμ΅μμ κ΄μ¬μ¬μ λͺ©νμ κΈ°λ°ν κ°μΈνλ μ°Έμ¬ κ²½λ‘λ₯Ό μ 곡ν μ μλ€. κ·Έλ¦¬κ³ λ€μν ννμ νμ νν(μμ±, ν
μ€νΈ, μμ, μ½λ)μ μμ©νκ³ , νμ μ€μ μ¬λΆκ° μλ νμ΅ μ±κ³Όλ₯Ό κΈ°μ€μΌλ‘ νκ°ν μ μλ€.
λ¬Έμ λ κΈ°κ΄λ€μ΄ AIλ₯Ό νμ©νμ¬ UDL μμΉμ μ§μ μ± μκ² κ΅¬νν κ²μΈμ§, μλλ©΄ μ€μ§μ μΈ λ΄μ© μμ΄ ν¬μ©μ μΈμλ§μ λ§λ€μ΄ λ΄λ λ° AIλ₯Ό μ¬μ©ν κ²μΈμ§μ μλ€.
μ₯μ νμμ μν UDL μ€ν
Wahyuni, Pantiwati, Sunaryo(2025)λ UDLμ΄ κ³ λ±κ΅μ‘μμ μ₯μ νμμ μν ν΅ν©κ΅μ‘μ μ΄λ»κ² κ°νν μ μλμ§ κ²ν νλ€. ν΅ν©μ κ³ λ±κ΅μ‘μ νμ΅μ λν 곡νν μ κ·Όμ 보μ₯νλ λ° νμμ μ΄λ€. UDLμ λ€μν νμ΅ μꡬλ₯Ό μμ©νκ³ ν΅ν©μ μ΄μ§νλ μ μ°ν κ΅μ νλ μμν¬λ₯Ό μ 곡νλ€.
PRISMA λ°©λ²λ‘ μ μ μ©ν μ΄ μ²΄κ³μ λ¬Έν κ³ μ°°μ, UDL μμΉμ΄ κ΄λ²μνκ² μ§μ§λ°κ³ μλ λ°λ©΄ μ€ν μ λ΅μ μ§μ μμ€κ³Ό λ²μμμ μλΉν μ°¨μ΄λ₯Ό 보μΈλ€λ μ μ λ°κ²¬νλ€. ν¨κ³Όμ μΈ UDL μ€νμ μν΄μλ λ€μμ΄ νμνλ€: κΈ°κ΄μ νμ (μ κ·Όμ±μ λ¨μν κΆμ₯νλ κ²μ΄ μλλΌ μ무ννλ μ μ±
), κ΅μμ§ μλ κ°λ°(UDL μμΉκ³Ό μ€μ μ μ©μ κ΄ν νλ ¨), κΈ°μ μΈνλΌ(λ€μν μμμ μ§μνλ νλ«νΌκ³Ό λꡬ), κ·Έλ¦¬κ³ μ§μμ μΈ νκ°(UDL μ€κ³κ° μ€μ λ‘ μ₯μ νμμ μ±κ³Όλ₯Ό ν₯μμν€λμ§ νκ°).
κ΅μ€μ λμ΄μ UDL
Gilleran Stephens, Antwi, Linnane(2025)μ UDLμ νκ²½κ΅μ‘ μμλ¦¬μΉ νλ‘κ·Έλ¨ μ¬μ€κ³μ μ μ©νλ©°, UDL μμΉμ΄ 곡μμ μΈ κ΅μ€ μμ
μ λμ΄ λΉκ³΅μ λ° λ¬΄νμ κ΅μ‘ λ§₯λ½μΌλ‘κΉμ§ νμ₯λ¨μ 보μ¬μ€λ€. UDLμ λ³΄λ€ ν¬μ©μ μ΄κ³ μν₯λ ₯ μλ κ²½νμ μ°½μΆνκΈ° μν νλ μμν¬λ‘ μ μ©λλ€.
UDLμ μμλ¦¬μΉ λ° κ³΅κ³΅ μ°Έμ¬μ μ μ©νλ κ²μ μ€μν μλ―Έλ₯Ό μ§λλλ°, μ΄λ UDLμ΄ κ³΅μ κ΅μ‘μλ§ κ΄λ ¨λλ€λ κ°μ μ λμ νκΈ° λλ¬Έμ΄λ€. κ³Όν νμ€ν°λ², μ§μμ¬ν μν¬μ, λ°λ¬Όκ΄ νλ‘κ·Έλ¨, κ³΅κ° κ°μ° λͺ¨λ λ€μν μ°Έμ¬, νν, νλ μλ¨μΌλ‘λΆν° μ΄μ μ μ»μ μ μλ€. μ΄ λ
Όλ¬Έμ νλ‘κ·Έλ¨μ μ λ©΄μ μΌλ‘ κ°νΈνμ§ μκ³ λ UDL μμΉμ μ¬μ©νμ¬ κΈ°μ‘΄ νλ‘κ·Έλ¨μ μ¬μ€κ³νλ λ°©λ²μ λν μ€μ©μ μΈ μ§μΉ¨μ μ 곡νλ€.
κ΅μ μλ κ°λ°: λΉ μ§ μ°κ²° κ³ λ¦¬
Dell'Anna, Marsili, Bevilacqua(2025)λ ν΅ν©μ μ¦μ§νλ λ° μμ΄ κ΅μ μλ κ°λ°κ³Ό UDLμ κ΅μ°¨μ μ κ²ν νλ€. λνμ κ· λ±ν κΈ°νμ μ κ·Όμ±μ μ¦μ§νλ λ° ν΅μ¬μ μΈ μν μ νλ©°, μ΄ λ
Όλ¬Έμ κ΅μ μλ κ°λ°(κ΅μλ², μ λ¬Έμ± κ°λ°, μ‘°μ§μ μ°¨μμ ν¬κ΄)κ³Ό UDLμ΄λΌλ λ κ°μ§ μ°κ΅¬ μμμ ν΅ν©νλ€.
μ΄ λ
Όλ¬Έμ κ΅μ μλ κ°λ° μμ΄λ UDL μ€νμ΄ μ€ν¨νλ€κ³ μ£Όμ₯νλλ°, μ΄λ μμ κ²ν ν κ΅μ¬ μ λ¬Έμ± κ°λ° λ¬Ένμ κ²°λ‘ κ³Ό μΌμΉνλ λ°κ²¬μ΄λ€. UDL μμΉμ λν νλ ¨μ λ°μμ§λ§ μ€μ μ μ© λ°©λ²μ λ°°μ°μ§ λͺ»ν κ΅μμ§μ μ΅μν κ΅μλ²μΌλ‘ νκ·νλ€. μΌνμ± UDL μν¬μμ λ°μμ§λ§ μ§μμ μΈ μ§μμ΄ μλ κ΅μμ§μ μ΄λ €μμ΄ μκΈ°λ©΄ μλ‘μ΄ μ€μ²μ ν¬κΈ°νλ€.
UDLλ₯Ό μν ν¨κ³Όμ μΈ κ΅μ κ°λ°μ λ€μμ νμλ‘ νλ€: μ§μμ μΈ μ°Έμ¬(μΌνμ± μν¬μμ΄ μλ), νλ¬Έ λΆμΌλ³ μ μ©(곡νμμμ UDLμ μΈλ¬Ένμμμ UDLκ³Ό λ€λ₯Έ μμμ λ€λ€), λλ£ νμ΅ κ³΅λ체(κ΅μμ§μ΄ μλ‘μ μ€ν κ²½νμΌλ‘λΆν° λ°°μ°λ κ²), κ·Έλ¦¬κ³ μ λμ μΈμΌν°λΈ(μΉμ§ λ° μ’
μ μ¬μ§κΆ κ²°μ μμ UDL λμ
μ μΈμ νλ κ²).
AI: UDLμ μ΄μ§μ
Mallary, Moore, McClain(2025)μ μ±μΈ λ° κ³μ κ³ λ±κ΅μ‘μμ UDLκ³Ό μμ±ν AIλ₯Ό ν΅ν©νλ κ²μ λ³νμ μ μ¬λ ₯μ νꡬνλ€. UDLμ μμΉμ μ½ν
μΈ μμ±, κ°μΈν, μ κ·Όμ± λΆμΌμμμ GenAI μλκ³Ό μ°κ³ν¨μΌλ‘μ¨, μ΄ λ
Όλ¬Έμ κΈ°μ κ°ν ν¬μ©μ κ΅μλ₯Ό μν νλ μμν¬λ₯Ό μ μνλ€.
AI-UDL ν΅ν© νλ μμν¬λ ꡬ체μ μΈ μλμ λ€μκ³Ό κ°μ΄ μ μνλ€:
- λ€μν νν μλ¨: AIκ° λ체 μ½ν
μΈ νμ(λμμμ ν
μ€νΈ μμ½, μ λμ λ°μ΄ν°μ μκ°μ νν, 볡μ‘ν ν
μ€νΈμ λ¨μνλ μΈμ΄ λ²μ )μ μλμΌλ‘ μμ±νλ€.
- λ€μν μ°Έμ¬ μλ¨: AIκ° νμ΅μμ κ΄μ¬μ¬, μ¬μ μ§μ, λκΈ°μ μ νΈλλ₯Ό λ°νμΌλ‘ νμ΅ κ²½λ‘λ₯Ό κ°μΈννλ€.
- λ€μν νν μλ¨: AIκ° κ³΅ν΅ νμ΅ μ±κ³Όμ λν΄ λ€μν νμ μ°μΆλ¬Ό(μμ± λ
Ήμ, μκ°μ λ°ν, μμ± μμΈμ΄)μ μμ©νκ³ νκ°νλ€.
μ΄ νλ μμν¬λ κ°λ
μ μΌλ‘ μ λ§νμ§λ§ μ€νμμ μ°λ €λ₯Ό μ κΈ°νλ€: AIκ° μμ±ν λ체 ννμ μλ³Έμ λμμ€λ₯Ό ν¬μ°©νμ§ λͺ»ν μ μκ³ , AI κ°μΈνλ μμΈλ νμ΅μμ μ νμ§λ₯Ό νμ₯νκΈ°λ³΄λ€ μ ννλ νΈν₯μ λ΄μ¬ν μ μμΌλ©°, λ€μν ννμ λν AI νκ°λ νμ¬μ AI μμ€ν
μ΄ μ΄λ €μμ κ²ͺλ λ§₯λ½μ νλ¨μ μꡬνλ€.
μ£Όμ₯κ³Ό κ·Όκ±°
<
| μ£Όμ₯ | κ·Όκ±° | νμ |
|---|
| UDLμ μ₯μ νμμ νμ΅ μ±κ³Όλ₯Ό κ°μ νλ€ | Wahyuni et al.(2025): κ·Όκ±°λ μ§μ§νλ μ€ν μ§μ΄ λ€μν¨ | β
μ§μ§λ¨(μ€νμ 쑰건λΆ) |
| UDL μμΉμ 곡μμ μΈ κ΅μ€ μμ
μ λμ΄ μ μ©λλ€ | Gilleran Stephens et al.(2025): μμλ¦¬μΉ νλ‘κ·Έλ¨μμ μ±κ³΅μ μ μ© | β
μ§μ§λ¨ |
| κ΅μ κ°λ°μ UDL μ€νμ νμμ μ΄λ€ | Dell'Anna et al.(2025): μ§μμ μΈ κ΅μ μ§μ μμ΄λ UDLμ΄ μ€ν¨ν¨ | β
μ§μ§λ¨ |
| AIλ UDLμ λκ·λͺ¨λ‘ ν¨κ³Όμ μΌλ‘ μ€νν μ μλ€ | Mallary et al.(2025): νλ μμν¬ μ μλ¨; κ²½νμ κ²μ¦ μ νμ | β οΈ λΆνμ€ |
μμ¬μ
UDLκ³Ό AIμ μλ ΄μ ν¬μ©μ κ΅μ‘μ μμΈκ° μλ κΈ°λ³Έκ°μΌλ‘ λ§λ€ κΈ°νλ₯Ό μ°½μΆνλ€. κ·Έλ¬λ μ΄ κΈ°νλ₯Ό μ€ννκΈ° μν΄μλ μ κ·Όμ± μ§μμ νλ°©νλ AI λꡬμ λ¨μ λμ
μ΄ μλ, κ΅μ κ°λ°, μ λμ μΈνλΌ, κ·Έλ¦¬κ³ μ§μμ μΈ νκ°μ λν ν¬μκ° μꡬλλ€. κΈ°μ μ UDL μμΉμ μ€νν μ μμ§λ§, κ·Έκ²μ μ€μ§ κΈ°μ μ΄ μμΉμ μ μνλ κ²μ΄ μλλΌ μμΉμ΄ κΈ°μ μ μ΄λ λλ§ κ°λ₯νλ€.
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), 380β397.
[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).