Challenge 3: Curriculum-Aligned Script Generation for Educational Videos

Short-form educational videos and animations are exploding in popularity across schools, online learning platforms, and social media. AI video-generation tools now allow anyone to turn written scripts into full visual explainers, but the quality of the final learning experience still depends almost entirely on the underlying script.

Most teachers and curriculum designers are not trained in writing instructional scripts that are:

• Pedagogically sound
• Curriculum-aligned
• Culturally safe
• Free from misconceptions or hallucinations
• Sequenced using proven learning science (e.g., retrieval, dual encoding, gradual release, interleaving)

As a result, AI-generated videos often contain oversimplified explanations, incorrect visual metaphors, distorted conceptual models, and culturally inappropriate representations, weakening learning and trust.

The output of the system is text only (e.g., scene-by-scene narration, captions, camera directions, visual description cues, accessibility descriptors, etc.). Another AI tool later converts this script into a video or animation.

Target users: Teachers, NGOs, curriculum designers, EdTech developers, and students who rely on AI for instructional content.

Possible solutions:

• Curriculum-aligned script builder
• Pedagogical-sequence frameworks (e.g., "I do / we do / you do", worked example → faded example)
• Cognitive-science-driven scene templates
• Hallucination-risk checking and factual-consistency safeguards
• Inclusive/culturally safe narrative patterns
• Adaptive scripts for different audiences (e.g., Year 3 vs Year 10)

• Use CurricuLLM-AU API to embed verifiable curriculum outcomes, concepts, and misconceptions guidance.
• Scripts should incorporate evidence-based teaching strategies:
  • Spaced retrieval
  • Multi-modal encoding
  • Worked/faded examples
  • Interleaving and contrastive explanation
  • Common misconception warnings
• Ensure scripts include clear direction for animation/video tools:
  • Scene descriptions
  • Character/voice guidance
  • Text overlays
  • Accessibility captions/ALT-style visual instructions
• UX must be teacher-friendly — the user describes an idea in natural language and receives a polished, ready-to-generate script.
• Cultural and contextual safety — avoid stereotypes and ensure accuracy for Australian curriculum, locations, communities, and cultural examples.

• A functional prototype or proof-of-concept system that helps educators generate curriculum-aligned, evidence-based video/animation scripts while reducing hallucination and factual risk. The output is text scripts only — not animations or videos.
• A 5-minute video and pitch deck explaining:
  • The design principles
  • How evidence-based learning is embedded into the scripting process
  • How curriculum alignment and hallucination-prevention mechanisms operate