Radio Play
Live Participatory Worldbuilding and Performance with Generative AI
A workshop for SLSA 2022 Reading Minds @ Purdue. West Lafayette, Indiana. October 7-8, 1:00-3:00 EDT.
«< STREAM LIVE »> Saturday October 8, 2022 3:00-3:30 EDT (1900-1930 GMT) (recording coming soon)
Description
With recent advancements in machine learning techniques, researchers have demonstrated remarkable achievements in text synthesis, understanding, and generation. This hands-on workshop introduces state-of-the-art transformer model (GPT-3) through a participatory, interactive event culminating in a live radio show and internet transmission. Participants will gain experience with Open AI’s GPT-3 as members of an AI writers room. We will discuss issues of liveness and serendipity; possibilities for human/non-human co-authorship; and relate these computational processes to human language, perception, and embodiment. Please bring a text you would like to experiment with!
Schedule
Friday, October 8 - Session 1
| Time | Activity |
|---|---|
| 1:00 | Introduction - Background and Setup |
| 1:15 | Activity: warming up our language engines (Discord). Discuss |
| 1:30 | Writer’s Room 1: Worldbuilding (small groups) |
| 2:00 | Reconvene: share out and decide on concept, themes |
| 2:10 | Writer’s Room 2: Developing scenes (small groups) |
| 2:50 | Reconvene: share out scenes, sign up for production. Other elements (sponsor messages, etc.) |
| 3:00 | assign “homework” and adjourn |
Saturday, October 9 - Session 2
| Time | Activity |
|---|---|
| 1:00 | Finalize Grid and Scripts, Interstitial Bits |
| 1:20 | Preproduction and wrangle audio assets, rehearse |
| 2:00 | Reconvene: check in |
| 2:10 | UPLOAD assets. Reconvene: Sequencing blocks, timing |
| 2:50 | Rehearsal and Sound Check |
Saturday, October 9 - Performance
| Time | Activity |
|---|---|
| 3:00 | Performance and Broadcast |
| 3:30 | Debrief |
| 4:00 | Adjourn |
Tools and Resources
- Radio Play Slides (google slides
- Discord Server [LINK REMOVED] (download discord)
- Radio Grids (google sheets)
- OpenAI GPT-3 overview beta.openai.com/overview
References
- OpenAI GPT-3: Language Models are Few-Shot Learners, 2020
- To familiarize yourself with GPT-3, sign up for the beta (openai.com/api), and work through these simple AI playground examples: https://beta.openai.com/examples.
- Chat: https://beta.openai.com/examples/default-chat
- Science Fiction Book List: https://beta.openai.com/examples/default-sci-fi-book-list
- Advertisement from Product Description: https://beta.openai.com/examples/default-ad-product-description
Episode 2 Team
Ash Smith - Story, Worlds, Speculative Design Lab, Johnny Carson for Emerging Media Arts, UNL - asheveryday.com
Robert Twomey - Machine Cohabitation Lab, Johnny Carson for Emerging Media Arts, UNL - roberttwomey.com
Jinku Kim - Johnny Carson for Emerging Media Arts, UNL - www.grayscale64.com
This workshop is made possible through generous support from the Johnny Carson Center for Emerging Media Arts