Happy Sunday! Here's what we've been working on.
"What would the Zootopia DMV scene look like in other movies?"
That question from my kids led to one of our coolest AI projects yet.
We took the iconic Flash the sloth scene and transformed it into four completely different movie styles: Toy Story, Inside Out, Finding Nemo, and Moana. Same scene composition. Same character positions. Completely different aesthetics.
The result? A stacked vertical video showing all five versions simultaneously. The kids loved it!
Here is a short video tutorial on how we did it:
How We Did It
The Process:
- Selected a simple 3-character scene (Judy and Nick at counter, Flash behind it)
- Used ChatGPT to generate detailed transformation prompts for each movie style
- Generated images with Nano Banana Pro (30-45 minutes)
- Transformed videos with Kling 01 (25 minutes)
- Stacked all 5 strips vertically in Premiere Pro (15 minutes)
Total time: 90 minutes first attempt. Future projects: 60 minutes.
Total cost: 240 Kling credits (8% of monthly allocation). About 12 projects possible per month.
What worked:
- Choosing scenes with clear character positions (service counters, tables, doorways)
- Detailed prompts specifying exact positioning ("LEFT, CENTER-LEFT, RIGHT")
- Using Kling 01 instead of Motion Control (accepts animated characters)
- Creating thin strip format (1920x384) for perfect stacking
What didn't work:
- Avatar transformation failed completely after 5 attempts (humanoid aliens too complex)
- Initial prompts created "hybrid" characters (fox features on Woody, rabbit ears on Jessie)
- Motion Control rejected animated characters
Full Tutorial Available Now on Website
The complete step-by-step guide is live on the site with:
- Exact ChatGPT prompt templates for any movie transformation
- Specific Premiere Pro positioning values (no guesswork)
- Credit budget planning (how many projects per month)
- Troubleshooting for common issues (hybrid characters, positioning problems)
- Why Avatar failed and what to avoid
Read the full guide: Transform Movie Scenes Across Different Styles
Why This Project Works for Families
Beyond the "cool factor," this project taught the kids:
- Visual style recognition across different animation studios
- How character design changes while maintaining the same scene
- AI prompt engineering basics (specific positioning language matters)
- That failures are part of the process (Avatar didn't work, and that's okay)
The conversation afterward was more valuable than the video itself. "How does the AI know what Woody looks like?" led to discussing training data, visual style consistency, and character design across different studios.
What's Next
Planning the next transformation project: The kids want to try The Lion King circle of life scene. Five animals instead of three characters. More complex. We'll see how it goes.
Also considering: Harry Potter scenes (castle transformations), Star Wars (different planet aesthetics), and Marvel (different comic book art styles).
The prompt structure is now saved. The workflow is tested. Time to see what else we can transform.
Quick Links:
Zootopia Transformation Tutorial
All Family Projects
Instagram @aidadlab
TikTok @aidadlab
[If you're on Instagram or TikTok, a follow would help the algorithm show this content to more families: @aidadlab on both platforms. Muchas gracias 🙏]
Thanks for following along on these experiments.
– Niall
