Skip to content

Assignment🔗︎

Goal🔗︎

This assignment lets you explore how deep generative models of human behaviour can be used to create a storytelling experience. You'll use machine-learning models (popularly called "generative AI") to script a story, turn it into spoken dialog audio, and animate it with gesturing avatars in a 3D environment. The focus is on getting a feel for these tools and how they can work together, using the files you generate and the parameters you control. Treat this as an opportunity to experiment, learn, and have fun with these technologies!

Story Theme🔗︎

Your story should revolve around a subject. The topics below are some ideas you can take inspiration from, but do not feel overly-constrained by them. You are free to explore at your own pace, so be creative!

Manipulation

Explore the power of persuasion gone right or wrong. Whether it’s useful health advice, emotionally charged therapy sessions, or political misinformation, this theme is all about bending the truth for better or worse.

Documentary

Dive into reality-based storytelling that unpacks complex topics like ethics, copyright, AI replacing human labor, or societal shifts. Think informative, reflective, and maybe even a little provocative.

Comedy

Make the audience laugh by telling something funny or cringy. Bring the laughs with awkward coworkers, alien neighbors, a stand-up gone wrong, spark a mockumentary mayhem or scramble-in some offbeat sci-fi skits.

Talent Show or Contest

Channel the chaos and charm of a live competition. Sing, joke, make weird sounds - whatever it takes to win over the crowd and maybe crash and burn doing it.

Freeform

No rules, no themes, now law or police to chase you. Just tell the story you want to tell, whether it’s heartfelt, absurd, poetic, or all of these combined. This is your blank canvas to draw on!

Process🔗︎

The assignment is split in three stages:

Stage 1. Group Work🔗︎

You work with your team to create text, audio, and animation files for your story.

  • Text (.txt): You create the script for your story with the help of LLMs like ChatGPT.
  • Audio (.wav): You use a text-to-speech model to generate speech audio. For extra fun, you are encouraged to use your own voices that you fine-tuned during the pre-assignment for the characters.
  • Animation (.bvh, .fbx): You use the WASP gesgen app to generate gesture animations from speech.

Click here for a detailed tutorial for the group work

Stage 2. Submission🔗︎

You package all created files to a .zip archive which you submit to us.

Click here for the submission instructions

Stage 3. Presentation🔗︎

We take your files and use them to produce movie sequences in Unreal Engine. We will put a brief intro and outro at the beginning and end of your story to glue your submission with the overarching presentation theme.

We then play these sequences in the Norrköping Visualization Center Dome. You will be invited to briefly present your team before we play your submision.

2024 Submissions (click to show)
2023 Submissions (click to show)
2022 Submissions (click to show)

Special Thanks🔗︎

The following great people helped us integrate the pre-assignment with the on-site assignment, including general support and coordination:

2025:

2024: