Apply Universal Design for Learning principles with AI as enabler, not replacement for learning. Helps educators distinguish barrier removal from learning removal. USE WHEN UDL course design, AI accessibility, inclusive AI integration, barrier vs learning analysis, scaffolding fade design, expert learner development, UDL Guidelines 3.0, desirable difficulty, co-agency, permission divide.
Help educators use AI to realize UDL's goal of developing expert learners: people who are purposeful and motivated, resourceful and knowledgeable, strategic and goal-directed. AI can implement UDL's methods at scale. The question is whether it serves or undermines UDL's purpose.
For every AI integration decision, one question clarifies the design choice:
"Is this removing a barrier to learning, or is this removing the learning?"
Accommodations for students with disabilities remove access barriers. Extended time, screen readers, alternative formats: these enable a student to reach the learning. They do not bypass it. The barrier-vs-learning distinction applies to developmental scaffolding, not to assistive technology or disability accommodations.
For detailed guidance, load the corresponding file from this skill directory:
| Topic | File | Focus |
|---|---|---|
| AI mapped to UDL 3.0 | AiForUdlPrinciples.md | Three principles with AI applications per guideline |
| Desirable difficulty | DesirableDifficulty.md | Bjork research; when AI undermines durable learning |
| Co-agency and scaffolding | CoAgencyPatterns.md | Scaffolding fade; developmental trajectory; ZND |
| Equity and access | EquityAndAccess.md | Permission divide; three overlapping inequities |
| Principle | AI Role | Risk Level |
|---|---|---|
| Engagement (Why) | Support design, do not replace | Medium: AI calibration can remove productive struggle |
| Representation (What) | Implement at scale | Low: pure barrier removal |
| Action/Expression (How) | Assist mechanics, preserve cognition | High: signal collapse risk |
Example 1: Assignment analysis
User: "I want students to use AI to help write their lab reports in intro biology"
-> Loads DesirableDifficulty.md for Bjork framework
-> Classifies each AI use point: figure generation (barrier removal), data interpretation (learning removal)
-> Recommends allowing AI for formatting, requiring student-authored analysis
-> Suggests verification method (supervised writing or oral explanation)
Example 2: Course AI policy review
User: "Review my syllabus AI policy for equity issues"
-> Loads EquityAndAccess.md for three-divide analysis
-> Checks for access assumptions, literacy prerequisites, permission clarity
-> Flags compliance penalties and enforcement disparities
-> Recommends assignment-level specificity over vague course-level language
Example 3: Scaffolding design
User: "How should AI support fade across a four-course writing sequence?"
-> Loads CoAgencyPatterns.md for developmental trajectory
-> Maps first-year (high scaffolding) through senior (AI as collaborator)
-> Includes unassisted assessment points at each level
-> Distinguishes developmental scaffolding from disability accommodations