Talk:Community Wishlist/W341

Caching videos on Commons

In theory, AI-generated videos would not need to be exhaustively stored on Commons. Videos could, instead, be cached on Commons as they could be re-generated, as needed, from screenplay or storyboard input data stored on Wikistoryboards. AdamSobieski (talk) 10:11, 14 December 2024 (UTC)

Generating individual panels from wikitext

While the precise nature of these "storyboards" or "screenplays" remains to be determined, in theory, individual panels of storyboards could each be generated from wikitext descriptions which could, thereafter, annotate resultant images in the panels. Resultant images could be but one component of a panel. AdamSobieski (talk) 10:48, 14 December 2024 (UTC)

Hypermedia

In theory, "storyboards" or "screenplays" could be hypermedia documents, e.g., capable of referring to other reusable "storyboard" panels or sequences via hyperlink. These could, then, form a web of documents. AdamSobieski (talk) 10:42, 14 December 2024 (UTC)

Enabling interactivity

While interactivity could be delivered to end-users by allowing navigation options between segments of storyboarded video (placing hyperlinks at the ends of segments), storyboards could, additionally, be interactive, capable of branching like choose-your-own-adventure stories. From these branching storyboards, interactive videos (films) could be generated and stored.

Individual and collections of interactive videos could support personalization and adaptivity with respect to educational content (navigation between segments could be based on audience members' performances on interspersed quizzes).

With respect to AI, other possibilities include that some AI systems could generate and evaluate "storyboards" or "screenplays" multimedia documents, while other AI systems could generate educational videos from these on-the-fly. AdamSobieski (talk) 10:21, 14 December 2024 (UTC)

Brainstorming, perhaps a new technology, a combination of screenplay, storyboard, and process model, could be of use.

As envisioned, "storyboards" would be modular and reusable, like computer-programming functions or template metaprograms.

As a function is a callable sequence of instructions, a "storyboard" could extend upon the concept of a callable sequence of panels. "Storyboards" could declare and utilize both input parameters and local variables.

Beyond simply hyperlinking to one another, "storyboards" could call one another, passing arguments to one another. These "storyboards" could have control flow and branching, resembling both functions and process models. AdamSobieski (talk) 12:08, 14 December 2024 (UTC)

Interoperability with artificial-intelligence agents via video tracks

In theory, generated (interactive) videos could utilize either timed-text or JSON/JSON-LD data in video tracks to interoperate with AI systems on either servers or users' devices.

Dialogue occurring in videos could be copied into AI systems' chat contexts, attributed to the characters in the videos.

Via such interoperability, users could ask other, interoperating AI assistants questions about events or dialogue in a video and their assistants could answer these questions.

Additionally, users could direct their questions to those characters in (interactive) videos; answering these could be accomplished by generating more (interactive) video content, in an on-the-fly manner. AdamSobieski (talk) 00:34, 15 December 2024 (UTC)

Wikipedia integration

Might computational screenplays be automatically generated from Wikipedia articles and collections of these articles? AdamSobieski (talk) 14:56, 15 December 2024 (UTC)

Sequential art and graphic novels

While the motivating use case involves supporting the creation and editing of "screenplay" or "storyboard" inputs for AI-generated single-character educational videos where a tutor would stand before a board or screen to perform instruction and to provide quizzes, another use case for the same software components would be a platform for the collaborative creation and editing of sequential art and graphic novels. AdamSobieski (talk) 11:00, 14 December 2024 (UTC)

Wikistories

This is similar to the existing project mw:Wikistories. * Pppery * _{it has begun} 17:06, 25 December 2024 (UTC)

Unclear

Could you explain what you mean by "computational storyboards" a bit more? A link there would also help. Maybe you mean scripts for animated and AI-generated videos. But then couldn't one put the script into the video description or on its talk page and put such videos all in one category… Prototyperspective (talk) 19:21, 21 November 2025 (UTC)

Yes, I was brainstorming about online spaces for collaboration on scripts, screenplays, or storyboards for AI-generated animations and videos. Goals of this community wishlist item include the democratization of creative writing, storytelling, comic-book creation, and animation and video production.

As for hyperlinks, there are, since this proposal started, a much larger number of existing products and services available online in the space of AI-enhanced story creation and multimedia content generation. So, people can, today, purchase a product or subscribe to a service to rapidly prototype, create, publish, and share their stories with one another. People can create and share creative content with one another more easily than ever before.

With respect to "computational" storyboards, I was brainstorming about how one might go about enabling the reuse of script, screenplay, or storyboard components, across projects. Need every project start from scratch? How might one creator or creative team reuse components or content from other projects? How might a creator or creative team create a story, comic book, animation, or video so that others could reuse those components from its creative process?

That is a good idea that any ingredients to generate a story, comic book, animation, or video would go into its talk page or some other accompanying page. AdamSobieski (talk) 20:45, 21 November 2025 (UTC)

I doubt many people are interested in recreating a video which was already implemented by somebody else though. A script is usually only used by one or so video and then not very useful anymore for new videos. Maybe the best thing to do here would be some concrete example where you put the things on Commons which would make this whole things and its potential usefulness and applications much more understandable imo.

For example, consider how few films there are that are very similar to an earlier book/… adoption. They usually vary a lot. Also videos and images can already be found on Commons and cropped or cut as needed. You could also create a template on Commons for when things are put onto the talk page. Haven't seen any case of this so far. Prototyperspective (talk) 21:05, 21 November 2025 (UTC)

Good points with respect to creativity and originality. Being able to search for, retrieve, and analyze existing content and process-related components can enhance the learning of new tools, techniques, and styles, in particular for students.

For a concrete example, when creating a screenplay and/or storyboard for a scene of a foot-chase through a forest, a (student) creator or creative team could search for existing visually similar scenes, view panels, notes, instructions for AI rendering systems (i.e., prompts), and so forth, and then create (and subsequently revise after viewing renderings) a storyboard sequence which suits their specific needs.

Another "computational" storyboarding topic, then, could be exploring how to enhance the indexing, search, and retrieval of and into sequences of panels, and enabling users to be able to watch corresponding video clips generated by AI systems for selected sequences of panels. Conversely, being able to navigate back from selections of videos to input storyboard sequences could be useful.

A user story, then, might include AI-equipped student filmmakers searching Commons for "open-source" storyboards, with portions of these storyboards mapping to selections of videos, which they could view to learn techniques, styles, and otherwise how to get their precisely desired results from varied inputs, including prompts or instructions to AI systems. AdamSobieski (talk) 22:06, 21 November 2025 (UTC)

Also, your idea to use talk pages or other accompanying pages to store and present the origin processes of AI-generated multimedia stored in Commons generalizes to include images and visualizations and infographics. Transcripts of man-machine multimodal conversations resulting in multimedia resources could be persisted (a form of provenance metadata?). These transcripts would also be educational, enabling learning by example with respect to using AI systems to generate multimedia. AdamSobieski (talk) 11:07, 22 November 2025 (UTC)