Multimodal Tailored Learning Support

What

Tool that helps participants understand context and information for deliberative processes through multiple formats (text, audio, video, VR, games), adapting to individual learning styles and measuring understanding non-intrusively.

Why

Increase learning rate ↑ allowing reduced process time ↓, improve accessibility ↑ for diverse participants, and cut learning costs ↓ while achieving better informedness outcomes ↑.

Problem Definition

Participants must quickly understand complex context for decisions, including issue-specific information (dynamics, options, tradeoffs, stakeholder views) and process context (goals, structure, outputs). Current approaches rely on 50-page documents of varying quality, expert presentations, and facilitated activities. This creates barriers for different learning styles, abilities, and time constraints. The one-size-fits-all approach leaves many participants inadequately informed, limiting deliberation quality.

Definition of Success

Achieve demonstrably better informedness: 30 minutes with the tool outperforms 2 hours of self-directed reading. Support three time scenarios: 5-minute overview, 2-hour standard, 15-hour deep dive. Cost <$5 per participant learning hour. For a Citizens’ Assembly (100 participants, 15 learning hours each): significant reduction in costs from ~$50k to ~$7.5k through automated content delivery and personalization. Enabling previously unfeasible learning support while improving outcomes across all participant demographics.

Requirements

• Content Delivery: multi-format output (text, audio, video, VR experiences, interactive tutorials, games); adaptive selection based on individual learning preferences; and dynamic content adjustment based on comprehension tracking.
• Quality Assurance: bias and accuracy guarantees with verification mechanisms; and source grounding and citation tracking,
• Conflict handling: present disagreements transparently without quality judgments; and logging and evaluation systems.
• Technical Implementation: non-intrusive understanding measurement; participant content contribution system; multiple interfaces (web, chatbot, voice assistant); and integration with existing deliberation platforms.

Existing Limitations

Current technology attempts include Custom GPTs (format-constrained, expensive, requires technical knowledge) and NotebookLM (text-heavy, non-interactive, only reformats/summarizes). Traditional approaches use static 50-page documents, one-way expert presentations, and resource-intensive facilitated activities. No existing solution combines adaptive multi-format delivery, personalization, and comprehension tracking. Quality of current materials varies widely with no standardization or outcome measurement.

Milestones

1. Demonstrate tool with measurable learning improvements (RCT preferred).
2. Achieve 2x learning speed increase validated through testing.
3. Deploy with practitioner group obtaining positive testimonials.
4. Scale to support full assembly with documented outcome improvements.

Starting Points

1. Read ‘Facilitating Deliberation: A Practical Guide by MosaicLab’ for process context.
2. Contact info@aidemocracyfoundation.org for practitioner partnerships.
3. Build on educational technology research in adaptive learning systems.
4. Leverage existing work in comprehension assessment and learning analytics.
5. Partner with upcoming assemblies for real-world testing.