Designing AI-Era Learning Experiences

Designing AI-Era Learning
with Mayer’s Principles, UDL, and High-Impact Practices

Photo by Kelly Sikkema on Unsplash

Introduction

This guide offers a practical blueprint for building online learning that is evidence-based, inclusive, and future-ready. It blends three powerful frameworks: Mayer’s multimedia principles for cognitive effectiveness, Universal Design for Learning (UDL) for equity and accessibility, and High-Impact Practices (HIPs) for deep, applied engagement. Together, they help teams create courses that are clear, motivating, and built for durable learning and transfer while thoughtfully leveraging AI where it adds value.

Theoretical Framework

Mayer’s (2001) multimedia principles show how words and images can be orchestrated to minimize extraneous load and maximize meaningful learning. Applied well, they move courses beyond presentation into active processing, retention, and transfer.

Universal Design for Learning (UDL) ensures courses work for all learners by proactively embedding multiple means of representation, engagement, and expression. UDL both reduces barriers and expands learner choice and agency (CAST, 2018).

High-Impact Practices (HIPs) emphasize authentic, collaborative, feedback-rich learning that builds persistence, belonging, and real-world readiness. Designing with HIPs keeps online experiences rigorous, reflective, and transformative (AAC&U, 2025).

Updating Multimedia Principles for the AI Era

Mayer’s foundations remain essential. What’s new is the chance to personalize and scaffold without losing clarity: AI can help declutter materials, offer alternative modes, provide formative nudges, and support multilingual access. This expansion dovetails with UDL’s emphasis on flexibility and equity and with HIPs’ emphasis on active, reflective, and applied learning. Equally important are guardrails: transparency about when/how AI is used; privacy and academic-integrity norms; and learner agency in choosing when to use AI support.

The 12 Revised Principles (with AI-era tactics)

For each principle, keep the classic core, then consider design-time and learner-side enhancements. Use AI as a progressive enhancement, not a requirement.

1. Coherence → Dynamic Coherence
   Classic: Exclude unnecessary words, pictures, sounds.
   Design with AI: Auto-summarize/declutter; provide “concise vs. extended” versions.
   Learner use: Let learners ask a GPT to condense, define terms, or expand examples without changing core content.
   
   
2. Signaling → Adaptive Signaling
   Classic: Use cues to highlight essentials.
   Design with AI: Insert data-triggered cues (“pause & try,” icons) based on pain points or analytics.
   Learner use: GPT tutors nudge reflection or practice when confusion is detected.
   
   
3. Redundancy → Learner-Controlled Redundancy
   Classic: Prefer narration + graphics over narration + graphics + on-screen text.
   Design with AI: Offer captions/transcripts/summaries as optional layers (auto-generated, human-checked).
   Learner use: Toggle captions/transcripts; request multilingual summaries on demand.
   
   
4. Spatial Contiguity → Extended Spatial Contiguity
   Classic: Place related words and pictures near each other.
   Design with AI: Auto-align labels/callouts near visuals; generate region-specific alt text.
   Learner use: Context-aware chat sits next to the exact diagram/frame.
   
   
5. Temporal Contiguity → Synchronized Temporal Contiguity
   Classic: Present corresponding words and visuals together.
   Design with AI: Sync narration, captions, highlights; auto-chapter long videos.
   Learner use: “Explain what happened at 03:14” prompts time-linked explanations.
   
   
6. Segmenting → Segmenting with AI Scaffolding
   Classic: Present learner-paced segments.
   Design with AI: Break into micro-units with adaptive reflection/branching and pause nudges.
   Learner use: GPT helps pacing, micro-goals, and targeted review plans.
   
   
7. Pre-training → AI-Supported Pre-training
   Classic: Provide key concepts up front.
   Design with AI: Auto-glossaries, prerequisite checks, and role-based primers.
   Learner use: Ask GPT for analogies or tailored pre-reads before heavier content.
   
   
8. Modality → Expanded Modality
   Classic: Words + pictures > words alone; favor audio narration with visuals over dense text.
   Design with AI: Generate alternative modes (podcast recap, infographic, interactive walkthrough) from one source.
   Learner use: Choose preferred mode (audio recap, checklist, sim).
   
   
9. Multimedia → Generative Multimedia
   Classic: Present words + pictures (with purpose).
   Design with AI: Generate purposeful diagrams/storyboards/data views (with clear learning intent).
   Learner use: Co-create images/charts with GPT, then critique/revise for accuracy.
   
   
10. Personalization → Relational Personalization
    Classic: Use conversational, learner-friendly style.
    Design with AI: Tune tone for inclusivity; adapt examples to roles/contexts while keeping essentials constant.
    Learner use: GPT/avatars tailor explanations to goals, building confidence and persistence.
    
    
11. Voice → Authentic Voice Principle
    Classic: Human voice generally outperforms synthetic for narration.
    Design with AI: Use expressive TTS/avatars for access needs; keep a consistent narrator; disclose AI-generated audio.
    Learner use: Offer voice options (pace/accent) with content parity.
    
    
12. Image → Presence with Purpose
    Classic: Speaker image isn’t inherently helpful.
    Design with AI: Use instructor video/agents only to model thinking, demonstrate processes, or give feedback.
    Learner use: Avatars simulate scenarios for safe practice with targeted feedback.

Cross-cutting practices

• Transparency & Ethics: Clearly label AI-generated/adapted assets; publish privacy and integrity norms.
  
  
• Agency & Co-Creation: Invite AI-supported brainstorming/drafting with critique, revision, and attribution norms.
  
  
• Data-Informed Feedback: Provide growth-oriented dashboards and formative nudges to guide next steps.
  
  
• Generative Collaboration: Make authorship expectations explicit when AI assists group writing or media creation.

Alignment Matrix (Mayer + UDL + HIPs → Practice)

Revised Principle	UDL Alignment	HIPs Alignment	From Principle to Practice
Dynamic Coherence	Reduces barriers via clear, adjustable content	Scaffolds authentic tasks with clarity	Design: Auto-summaries; concise/extended views. Learner: GPT to condense/expand terms.
Adaptive Signaling	Supports self-regulation & attention	Formative checks; reflective practice	Design: Data-triggered cues. Learner: GPT prompts to pause/try/reflect.
Learner-Controlled Redundancy	Multiple means of perception & expression	Agency in revision processes	Design: Optional captions/transcripts/summaries. Learner: Toggle/multilingual.
Extended Spatial Contiguity	Clarifies perception; reduces split attention	Analysis of real artifacts	Design: Co-located labels; alt text per region. Learner: Chat anchored to artifact.
Synchronized Temporal Contiguity	Timing supports comprehension	Demonstration aligned to practice	Design: AI-synced narration/captions/chapters. Learner: Time-linked explanations.
Segmenting w/ AI Scaffolding	Manages load; supports metacognition	Structured practice with feedback	Design: Micro-units + adaptive prompts. Learner: Plan pacing/review with GPT.
AI-Supported Pre-training	Builds background knowledge	Readiness for integrative tasks	Design: Auto-glossaries, primers. Learner: Analogies/primers on demand.
Expanded Modality	Multiple means of representation & engagement	Authentic, inquiry-based work	Design: Alternate modes (audio/infographic/sim). Learner: Choose preferred path.
Generative Multimedia	Purposeful visuals reduce extraneous load	Creation with critique	Design: Generate diagrams with clear purpose. Learner: Co-create and critique.
Relational Personalization	Belonging; responsive language	Mentoring; instructor presence	Design: Inclusive, role-tuned examples. Learner: Goal-aware explanations.
Authentic Voice	Transparent options; autonomy	Authentic dialogue & artifacts	Design: Consistent narrator; disclose AI voice. Learner: Pace/accent choice.
Presence with Purpose	Representation that supports engagement	Guided practice; modeling	Design: Video/avatars only when they add value. Learner: Avatar sims + feedback.
Transparency & Ethics*	Equitable, transparent design	Trust for collaboration	Design: Label AI; publish guardrails. Learner: Reflect and attribute AI use.
Agency & Co-Creation*	Multiple means of expression	Applied, participatory learning	Design: AI-supported creation tasks. Learner: Draft→critique→revise cycle.
Data-Informed Feedback*	Executive function & self-monitoring	Iteration, reflection (e.g., ePortfolios)	Design: Growth dashboards, nudges. Learner: Monitor progress, set next steps.

*Cross-cutting practices applied alongside all 12 principles.

Putting it to work: Now / Near / Next Roadmap

Use this phased roadmap to turn principles into practice without waiting on perfect tools. Start with Now (no-AI essentials that strengthen clarity, signaling, segmentation, and reflection) moves that are platform-agnostic and lift quality for every learner. When ready, layer Near enhancements that use lightweight AI (auto-summaries, multilingual captions, time-coded chapters, simple nudges) to expand access and feedback without re-architecting your course. In Next, scale personalization and practice with context-aware help at the point of need, adaptive pathways, avatar simulations, and learner-facing growth dashboards. Treat the phases as additive and mix-and-match: choose feasible items, pilot with clear success criteria, and document guardrails for privacy and integrity. The aim is steady, evidence-based improvement that honors Mayer’s principles, advances UDL, and embeds HIPs in everyday learning.

• Now (no-AI baseline): Tighten coherence; add explicit signals; provide optional captions/transcripts; break lessons into micro-segments with planned pause prompts; use checklists and worked examples; add reflection before/after practice.
  
  
• Near (light AI support): Auto-summaries; multilingual captions; time-linked video chapters; simple analytics-based nudges; role-tuned examples.
  
  
• Next (scaled AI & data): Context-aware help next to artifacts; adaptive practice pathways; avatar simulations with targeted feedback; growth dashboards with learner-controlled data views.

You don’t need cutting-edge tools to honor these principles. Start with clarity, accessibility, and authentic practice. When AI is available, use it to enhance, not replace, good design. The result is online learning that is cognitively sound (Mayer), universally accessible (UDL), and deeply engaging (HIPs) built for real learners in real contexts.

Course Design Checklist

The following is a sample comprehensive, modality-agnostic, checklist you can use for in-person, hybrid, or online learning. It aligns Mayer + UDL + HIPs and adds on-ground considerations (room setup, live captioning, materials) alongside digital ones. Use “(if applicable)” to skip items that don’t fit your context.

A) Foundations & Alignment (any modality)

• Outcomes are measurable and mapped to activities, media, and assessments.
• Each session/module opens with a brief overview + relevance hook (why it matters).
• Worked examples, concept maps, or demonstrations clarify complex ideas.
• Reflection points appear before and after practice (think-alouds, exit tickets, discussion).
• Workload expectations (time-on-task) are clear for all formats.
• Logistics are set: space/tech checked, sightlines/audio verified, backup plan ready (slides/handouts/QRs).

B) Mayer’s 12 Principles with AI-Era Tactics (Now / Near / Next)

1) Coherence → Dynamic Coherence

·        Now: [ ] Trim extraneous text/images; provide concise + extended versions (handout/slide notes).

·        Near: [ ] Auto-summaries for slides/handouts (human-checked).

·        Next: [ ] Personalization rules surface concise vs. extended per learner profile.

2) Signaling → Adaptive Signaling

·        Now: [ ] Headings, callouts, and verbal cues; clearly marked “pause & try” moments.

·        Near: [ ] Time-coded chapters in recordings; simple analytics/polling nudges.

·        Next: [ ] Data-triggered cues adapt based on performance/engagement.

3) Redundancy → Learner-Controlled Redundancy

·        Now: [ ] Don’t read slides verbatim; pair narration + visuals; provide notes separately.

·        Near: [ ] Optional transcripts/captions/summaries; printed/large-print alternatives.

·        Next: [ ] Learners set default display (captions, transcript pane, summary language).

4) Spatial Contiguity → Extended Spatial Contiguity

·        Now: [ ] Labels/callouts next to visuals (slides, whiteboards, handouts, lab setups).

·        Near: [ ] Region-specific alt text/annotations; doc-cam views aligned to diagrams.

·        Next: [ ] Context-aware help/chat anchored to the exact artifact or station.

5) Temporal Contiguity → Synchronized Temporal Contiguity

·        Now: [ ] Align narration with demonstrations or board work; avoid lag.

·        Near: [ ] Auto-chapter recordings; verify caption timing; clickable agenda.

·        Next: [ ] “Explain this moment” prompts return time-linked micro-explanations.

6) Segmenting → Segmenting with AI Scaffolding

·        Now: [ ] Chunk sessions into micro-segments with movement or pause prompts.

·        Near: [ ] Adaptive reflection checks at boundaries (clickers, polls, quick writes).

·        Next: [ ] Branching pathways recommend revisit/advance per learner data.

7) Pre-training → AI-Supported Pre-training

·        Now: [ ] Key terms/priors up front; quick diagnostic or warm-up.

·        Near: [ ] Auto-generated glossaries/role-based primers (reviewed).

·        Next: [ ] Dynamic primers adjust to background-knowledge signals.

8) Modality → Expanded Modality

·        Now: [ ] Pair words + pictures; use physical models/manipulatives where relevant.

·        Near: [ ] Alternate modes (audio recap, infographic, lab demo video, checklist).

·        Next: [ ] Learners choose preferred mode; system remembers preferences.

9) Multimedia → Generative Multimedia

·        Now: [ ] Each visual/prop/demo has a clear learning purpose tied to outcomes.

·        Near: [ ] Generate supportive diagrams/storyboards; instructor verifies fidelity.

·        Next: [ ] Learners co-create visuals and critique for accuracy (studio time, gallery walk).

10) Personalization → Relational Personalization

·        Now: [ ] Conversational, inclusive tone; examples reflect diverse contexts.

·        Near: [ ] Role-tuned examples by learner choice (track cards, stations).

·        Next: [ ] Goal-aware tutoring/agents tailor explanations and practice sets.

11) Voice → Authentic Voice Principle

·        Now: [ ] Consistent voice; mic use for audibility; avoid talking while facing away.

·        Near: [ ] Offer pace/accent options in recordings; disclose synthetic narration.

·        Next: [ ] Selectable voice options with content parity.

12) Image → Presence with Purpose

·        Now: [ ] Instructor presence (live or video) used to model thinking/process.

·        Near: [ ] Short demo clips; doc-cam for procedures; targeted feedback videos.

·        Next: [ ] Avatar simulations for role-play with standards-aligned feedback.

C) Universal Design for Learning (UDL)

Multiple Means of Representation

• Captions/CART or interpreters for live sessions (if applicable); transcripts for recordings.
• Key ideas available in more than one format (slide + handout; diagram + description; model + photo).
• Reading order, headings, and materials are screen-reader friendly; print alternatives available.

Multiple Means of Engagement

• Choice in topics/examples; relevance to authentic contexts.
• Predictable structure and visible progress cues (agenda, timers, checkpoints).
• Low-stakes practice with immediate feedback (clickers, whiteboards, short quizzes).

Multiple Means of Action & Expression

• Options to demonstrate learning (paper/prototype/presentation/video/diagram).
• Rubrics describe quality across modalities (clarity, evidence, accuracy, reasoning).
• Supports for planning/executive function (checklists, timelines, milestones).

D) High-Impact Practices (HIPs)

• Authentic tasks mirror real problems, audiences, data, or clients.
• Frequent, substantive feedback (instructor, peer, automated formative).
• Significant time-on-task with staged milestones and revision cycles.
• Structured collaboration (roles, norms, equitable contribution checks).
• Diverse/global perspectives integrated into cases/examples.
• Threaded reflection (pre/during/post; learning journals; exit tickets).
• Culminating product (showcase, ePortfolio, poster session, demo day).

E) Assessment, Integrity, & Transparency

• Outcomes ↔ activities ↔ assessments are explicitly aligned.
• Rubrics cover content knowledge and process/communication.
• Clear AI use policy (allowed/restricted/prohibited) with examples.
• Citation/attribution norms for AI-assisted work; process evidence when needed.
• Integrity strategies: unique prompts, oral checks, versioned drafts, studio critiques.

F) Accessibility, Safety, & Logistics (Digital + Physical)

• Room & tech check: sightlines, audio, microphones, assistive listening systems.
• Lighting/contrast adequate; do not rely on color alone to convey meaning.
• Alt text/long descriptions for images/figures; tactile/large-print options (if applicable).
• Media players support captions, transcripts, and playback speed.
• Wayfinding & seating: accessible routes, reserved seating as needed.
• Downloadable or printed equivalents for interactive activities when feasible.
• Safety protocols for labs/fieldwork clearly briefed and posted.

G) Data-Informed Feedback & Analytics

• Formative checks inform next steps (reteach, extension, office hours).
• Learners can view progress and set goals (“next best action”).
• Instructor has a light-touch triage view (attendance, polls, quiz items).
• Data used to improve design, not surveil; privacy noted.

H) Ethics, Privacy, & Disclosure (AI & Recording)

• Label any AI-generated/adapted media; human review for accuracy/fairness.
• Plain-language privacy/recording notice; opt-out alternatives when feasible.
• Avoid uploading personally identifiable or protected data in tools.
• Provide non-AI paths for essential learning if AI is restricted.

I) Now / Near / Next Plan (per course, unit, or workshop)

• Now (no-AI baseline) items implemented (coherence, signaling, segmentation, reflection).
• Near pilots chosen with success criteria, small cohort, and timeline.
• Next roadmap drafted with dependencies (policy, platform, budget).
• Owner + due date tracked for each item.

J) Pre-Launch & Continuous Improvement

• Peer review against this checklist; issues resolved.
• Usability check with a few learners on flow and clarity (in person or remote).
• Accessibility spot-check (screen reader + keyboard only; live CART test if used).
• Pulse survey includes items on clarity, cognitive load, relevance, and access.
• Regular review cycle (e.g., each term) to refresh materials and data rules.

References:

Association of American Colleges & Universities (AAC&U). (2025). High-impact practices. https://www.aacu.org/trending-topics/high-impact-practices

CAST. (2018). Universal Design for Learning guidelines version 2.2. http://udlguidelines.cast.org

Mayer, R. E. (2001). Multimedia learning. Cambridge University Press.

National Standards for Quality Online Learning (NSQOL). (2025). National standards for quality online courses. https://nsqol.org/national-standards-for-quality-online-learning/