Some Brainstorming and Project Ideas

Community
, ,
1

Inspired by the GPT-4o event and technology update, I would like to share some brainstorming and project ideas.

When I used to play tabletop RPGs, my friends and I would often draw our characters, items, costumes, equipment, creatures, monsters, villains, battles, settings, and situations from play.

We also made graphic novels – by hand, the old-fashioned way. It was a time-consuming hobby and, from my perspective, having to draw each panel manually detracted from rapidly authoring and sharing stories.

Project idea 1: Uses of generative AI for creatives to make and share graphic novels.

Project idea 2: GMs can, today, use generative AI to rapidly prepare artistic resources, before sessions, to share with players during play. In the near future, AI assistants (provided with input materials from GMs, players, RPG vendors, adventure vendors, etc.) could listen to groups’ play sessions to provide them with real-time generated (and subsequently editable) multimedia content.

GM: Together, you march through the dark forest, pressing forwards into the darkness, brushing aside debris, and then, through an opening, you stumble upon a clearing… and a green dragon! (presses button on laptop)

AI Assistant: (generates and displays a 2D/3D photorealistic picture of the players’ characters, from behind, in a clearing surrounded by woods, facing a tremendous green dragon; camera slowly pans, revealing 3D effects)

Players: Cool!

Project idea 3: While there are AI-enhanced tools for character sketching, co-authoring scripts, and co-creating storyboards for (linear) stories (e.g., Boords), and, while there are tools for designing interactive stories (e.g., Twine), I could not readily find AI-enhanced tools for co-authoring, analyzing, and illustrating forms of interactive stories such as choose-your-own-adventure stories, digital gamebooks, or choose-your-own-adventure graphic novels.

Project idea 4: A new approach for co-creating stories with AI involves human authors or authoring teams as GMs for groups of designed AI players/characters. Recordings and transcripts from resulting adventures could be used to produce linear stories, illustrated stories, graphic novels, and/or videos.

Project idea 5: Building upon idea #4, adding instructional components, analysis and evaluation components, we can envision software for GM and/or improvisational theatre practice and training, also using AI players/characters.

Any thoughts on these ideas? Any (related or unrelated) project ideas to share?

1 Like
2

I had this idea for an ai choose your own adventure. Start with input of things like genre, setting, inspiration from other media sources, MC information like gender, appearance, background etc.

The user can then pick the story length and the AI generates a story but where the MC may make an important decision the user can make a choice. Eg MC is the mayor of a town and ends up saving it from bandits by convincing them to become soldiers.

The story could MC could be saved from one story and reused in another. Maybe choose to have all books created in the same universe/world.

3

Interesting story ideas.

If you like writing CYOA stories, you might enjoy the following paper:

Mawhorter, Peter, Michael Mateas, Noah Wardrip-Fruin, and Arnav Jhala. “Towards a theory of choice poetics.” In In Proceedings of the 9th International Conference on the Foundations of Digital Games . 2014. [PDF]

It discusses the poetics of choice points in branching stories, how choice points in these stories can contribute to creating intended literary and aesthetic effects for readers. Each path through a branching story is a literary experience for its readers and their decision-making processes at choice points are also a part of these experiences.

With respect to large language models (e.g., GPT-3) and co-creating, co-authoring interactive branching stories for existing tools (e.g., Twine), I more recently found:

Calderwood, Alex, Noah Wardrip-Fruin, and Michael Mateas. “Spinning Coherent Interactive Fiction through Foundation Model Prompts.” In ICCC , pp. 44-53. 2022. [PDF] [GitHub]

Our tool allows users to author interactive-fiction narratives with the Twine framework, alternately “passing the keyboard” between the system’s language model and the user at will.

It’s exciting to consider how advancing AI technologies, e.g., GPT-4 and GPT-4o, could be useful along these lines.

4

Hey Adam, Regarding Project Idea 1 check out The Simulation by Showrunner. Quote from a news article:

" The Simulation (formerly Fable Studio) announced 10 TV shows (web series) made with Showrunner, the company’s text-to-episode system, each in different styles, from anime to Pixar-style, to the cutout style of South Park.

“It’s the Netflix of AI,” founder and CEO Edward Saatchi told me. “Watch an episode, or make an episode.”
"

So let’s see, project idea 1 is to use of generative AI for creatives to make and share graphic novels. It seems that the above example gives text to video netflix style episodes. Even if it hallucinates a bit editing will be possible. So I think for your Project Idea 1 it’s already happening and Yes it’ll be a product to make creatives of any and all types.

1 Like
5

This might actually be happening already or will happen in the near future. ChatGPT shocked the world in Nov. 2022. It keeps shocking the world. Enterprises in the tech industry will have already seen its potential so I think game developers like Epic Games must have gotten their hands on AI creation systems to spin up environments rapidly as you said.

In due time a startup will get funding and will implement this. Ideation in AI startups is at a peak right now and I think very soon some funded startup(s) will develop this as well.

Check out what Qualcomm has said about education here:

It’s very clear that education will soon involve AI + VR. So it’s only a logical next step that authoring systems become AI + VR enhanced as well. Qualcomm is the chipmaker here. In this article the picture is very interesting. If you see the second figure in the article you’ll see a few students wearing VR headsets but they havent put them on their eyes yet. They look at a physical object and then put on their VR headsets to seemigly play with it in virtual reality environment. The whole article proposes (or predicts) a generative AI + VR settings. Qualcomm is therefore making chips which will make this happen within VR headsets. Generative AI will be there inside Virtual Reality environments. So Idea 4 and Idea 5 are definitely coming.

6

به من کمک می‌کنی تا اپلیکیشن شبکه اجتماعی که طراحی کردم را به شرکت ٱپنai دهیم