Overview

Playbook 02: AI Video Making — Kids Educational Media with Google Generative AI

Playbook: PB-02 (AI Video Making: Educational Kids Media & Language Learning)
Document Type: Domain-Specific Steering Guide & Authoring Standard
Target Audience: Year 1 Computer Science & Software Engineering Students Applicability: Mandatory for all research, authoring, and auditing agents working on PB-02 chapters
Core Tooling Stack: Google Generative AI Ecosystem (Gemini 2.5 Flash / Pro, Google Veo 2 / Veo 3, Imagen 3, Google Cloud TTS Journey / Studio Voices, Lyria / MusicFX, FFmpeg 7.0+)
Repository Target: playbooks/02-ai-video-making/

1. Mission & Domain Focus

The mission of Playbook 02 is to provide a comprehensive, production-grade technical engineering manual for creating high-engagement, pedagogically sound educational videos for children (with a primary focus on early childhood language learning, vocabulary acquisition, phonics, and storytelling) using the Google Generative AI Media Stack.

Why Educational Kids Video is Unique & Demanding

Creating video for children (ages 2–8) is fundamentally different from commercial cinema:

1. Pedagogical Pacing & Cognitive Load: Young language learners require deliberate pauses, clear visual-audio word association (the "dual-coding theory"), and structured call-and-response repetition cycles.
2. Visual Aesthetics & Anti-Uncanny Valley: Children reject hyper-realistic human faces that exhibit subtle video diffusion flaws. The ideal visual language consists of charming, high-appeal 3D cartoon, claymation, or storybook animation (Pixar/Disney/Aardman aesthetic) with warm, vibrant color palettes.
3. Mascot & Character Attachment: A recurring, emotionally expressive animated mascot (e.g., an inquisitive bear, a cheerful robot, an energetic bunny) drastically increases engagement and language retention.
4. Phonetic Clarity & Audio Pacing: Speech must be crisp, cheerful, perfectly enunciated at 80–90% standard speed, with exact synchrony between spoken phonemes and on-screen visual flashcards.

2. Mandatory Tool Currency & Horizon Review Protocol

⚡ Tool Currency Mandate: Authors and auditing agents must verify and utilize the latest generation of Google tools and model IDs. Never rely on deprecated or prior-generation model names.

Active Production Tool Matrix (2026 Edition)

3. The Google Generative AI Media Architecture

┌────────────────────────────────────────────────────────────────────────┐
│                   GEMINI 2.5 FLASH (Multimodal Director)               │
│ - Curriculum design (CEFR Pre-A1/A1 vocabulary, age-appropriate themes) │
│ - Shot-by-shot storyboards with visual, dialogue & audio cue tags      │
└───────────────────────────────────┬────────────────────────────────────┘
                                    │
       ┌────────────────────────────┴────────────────────────────┐
       ▼                                                         ▼
┌───────────────────────────────┐        ┌───────────────────────────────┐
│     IMAGEN 3 (Visual Master)  │        │   GOOGLE CLOUD TTS (Voice)    │
│ - 3D cartoon mascot turnaround│        │ - Journey & Studio voices     │
│ - High-contrast teaching keys │        │ - SSML prosody, pause control │
│ - Phonics & flashcard assets  │        │ - Phonetic enunciation tuning │
└──────────────┬────────────────┘        └───────────────┬───────────────┘
               │                                         │
               ▼                                         ▼
┌───────────────────────────────┐        ┌───────────────────────────────┐
│   GOOGLE VEO 2 / 3 (Video)    │        │     LYRIA / MUSICFX (Sound)   │
│ - Spatio-temporal DiT video   │        │ - Catchy mnemonic melodies    │
│ - Playful mascot animation    │        │ - Educational Foley & stingers│
│ - Squash-and-stretch physics  │        │ - Cheerful nursery rhythms    │
└──────────────┬────────────────┘        └───────────────┬───────────────┘
               │                                         │
               └────────────────────┬────────────────────┘
                                    │
                                    ▼
┌────────────────────────────────────────────────────────────────────────┐
│                  AUTOMATED POST-PRODUCTION (FFmpeg 7.0+)               │
│ - Bouncy karaoke-style vocabulary subtitles (syllable color highlight) │
│ - On-screen visual flashcards with popping animation                   │
│ - Automated audio ducking (-12dB background music when mascot speaks)  │
└────────────────────────────────────────────────────────────────────────┘

4. The 7 Universal Quality Acceptance Gates

Every chapter must satisfy all 7 acceptance gates before publication:

1. Gate 1: Zero Fluff & Pedagogical Engineering Rigor: Open directly with learning mechanics, cognitive load models, model parameters, and concrete generation architectures.
2. Gate 2: Mandatory Naive vs. Production Contrasts: Contrast amateur approaches (generic adult prompts, terrifying hyper-realistic faces, chaotic pacing) with child-friendly production standards.
3. Gate 3: Latest Google Model Configurations: Detail exact parameter dictionaries for Google Veo 2 / 3, Imagen 3, Gemini 2.5 Flash, and Cloud TTS Journey voices.
4. Gate 4: Quantitative Trade-Off Matrix: Benchmark generation latency, API token/render costs, aesthetic consistency, and child engagement scores across configurations.
5. Gate 5: The 10 Operational Failure Modes in Kids AI Video: Address edge cases like creepy uncanny-valley character drift, scary morphing, over-stimulating rapid cuts, and voice enunciation slurring.
6. Gate 6: Mandatory Hands-On Lab (Interactive Challenge): Every chapter must feature an explicit, step-by-step challenge where Year 1 students can design, script, or configure a real educational video asset.
7. Gate 7: Mandatory Recommended Answer & Executable Solution: Every lab must be accompanied by the official Recommended Answer—providing fully runnable, zero-dependency Python 3.11+ code or battle-tested Google prompt/SSML recipes targeting Gemini 2.5 Flash and Veo 2 with automated verification assertions.

5. The Master Syllabus

6. Pedagogical Directing Rules for Kids AI Media

1. The 3-Second Cognitive Rule: Never introduce a new word while the background or camera is moving rapidly. The camera must rest or slow down when the mascot introduces the core vocabulary word.
2. The Triple-Reinforcement Pattern:
   • Audio: Mascot says the word clearly ("Apple!").
   • Visual: Mascot points to the rendered 3D object (A red, shiny apple).
   • Text: High-contrast, bold, bubbly letters pop up on screen (A - P - P - L - E).
3. The 2.5-Second Repetition Window: When the mascot asks "Can you say Apple?", leave an exact 2,500ms silence in the audio track so the child viewer can say the word out loud before the mascot celebrates ("Great job!").
4. Zero-Uncanny Valley Policy: Strictly enforce stylized cartoon, claymation, or storybook aesthetics. Never use realistic human faces for toddler/young child content to prevent psychological discomfort.