Remember that hilariously awkward, cringe scene in Napoleon Dynamite (2004) where Napoleon gives Trisha a pencil drawing of herself, and she was completely speechless at how awful it was?
Well, thanks to ChatGPT (or any alternative, powerful AI image generator), you can now recreate that movie scene with yourself as Trisha. And yeah, that includes Napoleon’s absolutely horrendous drawing skills.
I decided to give this a shot using ChatGPT’s image generator, and the results were pretty damn funny and shockingly accurate.
This might just be the cringiest AI tutorial that you’ll find on the internet, but that’s what we’re here for.
Here’s how to do it yourself.
Let’s party.
How to Recreate the Napoleon Dynamite Drawing Scene With AI
The goal is to generate two separate images from the movie, but with your face replacing Trisha’s.
The first image is Napoleon’s terrible pencil drawing of you, complete with your name at the top. The second is you standing in Trisha’s living room, looking absolutely disgusted by the masterpiece you’re holding.
Here are the original scenes from the movie:
Original scene: Trisha reacting to Napoleon’s drawing

Original scene: Napoleon’s drawing of Trisha

Step 1: Upload Your Reference Images
Open ChatGPT and start a new conversation.
Upload three images:
- A screenshot of Napoleon’s original drawing of Trisha (hint: download the image from this article).
- A screenshot of Trisha reacting to the drawing in her living room (hint: download the image from this article).
- A clear photograph of yourself (or whoever you want to put in the scene).
For the best results, use a high-quality photo where your face is clearly visible. This will help the AI reproduce your facial features in both images.
Step 2: Copy and Paste This AI Prompt
AI assisted me in creating the following prompt to recreate both scenes as faithfully as possible, including the original living room, clothing, facial expressions, and Napoleon’s complete lack of artistic talent.
The prompt is quite long, but that’s intentional. There are a lot of details to get right, okay?
Copy and paste the entire prompt into ChatGPT:
Recreate two iconic scenes from the 2004 film Napoleon Dynamite, replacing Trisha with the person in my uploaded photograph.
YOUR TASK:
Generate TWO SEPARATE, photorealistic images in a single request.
Image 1: Napoleon's hilariously terrible pencil portrait.
Image 2: The same person reacting with disgust to that portrait in Trisha's living room.
These are TWO DIFFERENT SCENES, not alternative versions of one image.
REFERENCE IMAGES:
Reference 1: Napoleon's pencil drawing of Trisha.
Reference 2: Trisha reacting to the drawing with her mother behind her.
Reference 3: The person whose face will replace Trisha's.
Use Reference 3 to establish the person's facial identity in BOTH generated images.
==================================================
IMAGE 1: THE TERRIBLE PENCIL PORTRAIT
==================================================
Recreate Reference 1 with near-identical composition.
FORMAT:
16:9 cinematic aspect ratio.
COMPOSITION:
- Close-up of white paper attached to slightly larger brown cardboard.
- Two hands holding the paper near its lower edges.
- Same paper angle, camera perspective, and framing.
- Blue denim sleeve visible on the right.
- Softly blurred brown interior background.
- Warm, slightly faded early-2000s movie cinematography.
THE DRAWING:
Replace Trisha's portrait with an intentionally awful pencil drawing of the person in Reference 3.
CRITICAL: The drawing must replicate the SPECIFIC artistic incompetence of Napoleon's original portrait.
It must feature:
- Absurdly elongated, narrow facial proportions.
- Uncomfortably large, mismatched eyes.
- Awkwardly placed eyebrows.
- A badly drawn, narrow nose.
- A small, crooked mouth.
- An excessively long, angular chin.
- Poor facial symmetry.
- Strange shading that makes the face look unintentionally creepy.
- Dense, scratchy graphite lines.
- Amateur crosshatching.
- Uneven pencil pressure.
- Poorly rendered hair.
- Bizarre anatomical proportions.
Adapt the uploaded person's actual facial features, hairstyle, facial hair, and distinguishing characteristics into this terrible drawing style.
For example, if the person has a beard, exaggerate it with clumsy, scratchy pencil strokes. If the person has short hair, preserve that hairstyle but render it terribly.
IMPORTANT:
The portrait should NOT look like a skilled artist intentionally creating a funny caricature.
It must look like a genuinely untalented teenager sincerely attempting a realistic pencil portrait and failing spectacularly.
Preserve enough recognizable features that viewers can identify the person in Reference 3.
TEXT:
Replace "TRISHA" at the top of the paper with the person's first name.
Render the name in the exact same amateur, hand-drawn, three-dimensional block-letter style as the original.
If the person's name isn't supplied, ask for it before generating.
The finished image should look like an authentic screenshot from Napoleon Dynamite with only the portrait's subject and name changed.
==================================================
IMAGE 2: THE DISGUSTED REACTION
==================================================
Recreate Reference 2 with near-identical composition, except the person in Reference 3 replaces Trisha.
FORMAT:
4:3 aspect ratio.
CAMERA AND COMPOSITION:
- Same camera height and perspective as Reference 2.
- Medium shot of the main subject standing slightly left of center.
- Subject holding the terrible portrait with both hands.
- Paper positioned at approximately waist/chest height.
- Head tilted downward, examining the portrait.
- Same framing, body proportions, and positioning as the movie screenshot.
SUBJECT:
Replace Trisha's face with the person from Reference 3.
Preserve their actual:
- Facial structure.
- Eye shape and color.
- Nose.
- Mouth.
- Hairstyle and hairline.
- Facial hair, if applicable.
- Other recognizable facial characteristics.
Maintain an unmistakable likeness to the uploaded photograph.
CLOTHING:
Keep Trisha's original outfit:
- Blue denim jacket with oversized collar.
- White pants.
Adapt the clothing naturally to the person's body.
EXPRESSION:
This is the most important element of Image 2.
Reproduce Trisha's EXACT reaction to Napoleon's drawing.
The person's face should communicate disbelief, discomfort, confusion, and mild disgust.
Specifically:
- Furrowed eyebrows.
- Slightly wrinkled nose.
- Narrowed eyes.
- Tight, awkward grimace.
- Downward gaze fixed on the drawing.
- Subtle, deadpan disgust.
The expression must match Trisha's understated reaction, not an exaggerated cartoon expression.
The intended reaction is:
"What the hell is this supposed to be?"
BACKGROUND:
Reproduce the original living room from Reference 2 as faithfully as possible.
KEEP THE ORIGINAL MOTHER:
- Same woman standing in the background.
- Same facial appearance and hairstyle.
- Same purple outfit with pink accents.
- Same body positioning.
- Same facial expression.
- Same location relative to the main subject.
KEEP THE ORIGINAL HOUSE:
- Cream-colored leather sofa on the right.
- Patterned sofa on the left.
- Red pillow.
- Wooden archway leading into the kitchen.
- Brown wooden kitchen cabinets.
- Ceiling fan and overhead lighting.
- Wall decorations.
- Beige carpet.
- Original furniture arrangement.
- Same warm indoor lighting.
The mother and background should be slightly out of focus, exactly as in the reference screenshot.
Maintain the awkward, ordinary suburban atmosphere of the original film.
CRITICAL CONTINUITY:
The paper held by the person in Image 2 must be the SAME terrible drawing generated in Image 1.
Although the BACK of the paper faces the camera in the original reaction screenshot, it should clearly be the same cardboard-backed portrait from Image 1.
Do not change the person's face between images.
==================================================
VISUAL STYLE FOR BOTH IMAGES
==================================================
Recreate the authentic cinematography of Napoleon Dynamite (2004).
Both images must feature:
- Photorealistic live-action movie appearance.
- Early-2000s independent-film photography.
- Slightly faded, warm colors.
- Subtle film grain.
- Natural household lighting.
- Mild image softness.
- Accurate depth of field.
- Ordinary, unglamorous suburban environments.
- Awkward, deadpan visual comedy.
ABSOLUTELY AVOID:
- Modern cinematic lighting.
- Glossy commercial photography.
- AI beauty enhancements.
- Professional-looking pencil artwork.
- Digital illustration effects.
- Cartoon or anime aesthetics.
- Exaggerated comedic expressions.
- Changes to the original room.
- Changes to the mother.
- Unnecessary creative reinterpretation.
==================================================
MANDATORY GENERATION WORKFLOW
==================================================
You must perform TWO distinct image generations.
GENERATION 1:
Create Image 1, the terrible pencil portrait.
Use References 1 and 3.
Output one complete 16:9 image.
GENERATION 2:
Immediately create Image 2, the disgusted reaction.
Use References 2 and 3, plus your newly generated Image 1 for visual continuity.
Output one complete 4:3 image.
IMPORTANT:
Do not generate another variation of Image 1 during the second generation.
Image 2 MUST feature the person physically standing inside the living room, examining the paper with disgust, while the mother watches from behind.
Generate both images before considering the task finished.
DO NOT:
- Combine them into a collage.
- Place them side by side.
- Generate two drawings.
- Generate two reaction shots.
- Stop after the first image.
- Substitute a text description for either image.
If your image-generation system cannot reliably execute two generations in one call, execute two consecutive image-generation calls automatically.
Do not ask me to type "continue" between generations.
==================================================
FINAL QUALITY REQUIREMENTS
==================================================
Prioritize the following:
1. Faithfulness to the original movie screenshots.
2. Accurate likeness to the person in Reference 3.
3. Faithful reproduction of Napoleon's terrible drawing technique.
4. Accurate recreation of Trisha's disgusted expression.
5. Identical clothing, mother, furniture, lighting, and framing.
6. Visual consistency between the two generated images.
FINAL DELIVERABLE:
Exactly TWO SEPARATE images:
IMAGE 1:
The person's hilariously awful pencil portrait, held by two hands, with their first name written above it.
IMAGE 2:
The same person wearing Trisha's denim jacket and white pants, reacting with disgust to that portrait in the original living room, with the original mother standing behind them.
The results should look like two authentic, consecutive screenshots from Napoleon Dynamite, with the person in Reference 3 seamlessly replacing Trisha.
BEGIN GENERATING BOTH IMAGES NOW. Step 3: Check Out Your Results
Here’s what ChatGPT generated when I used my own photograph.
My Napoleon Dynamite pencil portrait:

My reaction to Napoleon’s artistic masterpiece:

I think Napoleon really captured my good side.
If your results don’t look quite right, ask ChatGPT to make some adjustments and/or upload additional images of yourself. You might need to correct your facial features, the drawing itself, or the reaction portrait.
And if ChatGPT only generates one image, simply ask it to generate the second scene using the first image as a reference.
Now, here are both before and after images:




Wrapping It Up
Now you can put yourself into movie scenes with near-zero effort. Another completely unnecessary but highly entertaining use case for AI image generation.
Personally, I think this is one of those AI image prompts that works especially well with photos of your friends or your significant other. Send someone a terrible Napoleon Dynamite portrait of themselves and see how they react.
Don’t stop with just doing Napoleon Dynamite. You can do this for any movie scene you’d like. And if you want to get really crazy, you can use an AI video generator to reproduce the scene in an even more realistic way.
Give it a try, and let me know how your images turn out in the comments section below.
Until next time, remember to run the prompts and prompt the planet.
