Conversation

but casual creator design patterns aren’t universally applicable. in particular, making the creative process totally frictionless isn’t always ideal when you want the user to slow down and think more deeply about what they’re doing & why – in other words, to facilitate reflection
1
2
13
and this is where reflective creators come in! we propose a set of preliminary design patterns for autotelic creativity support tools that strategically: - add friction - slow down the creative process - act as a *trusted adversary* to the user, like an editor to a writer
1
2
22
these design patterns draw inspiration from a wide range of existing human/machine co-creative systems: things like ’s SkyKnit system, our flagship example of *reflective repair* (generating half-viable artifacts for the user to fix)
Quote Tweet
Replying to @jonathanfly and @drbeef_
a few of us at UCSC have taken to calling this the "SkyKnit effect" (as in aiweirdness.com/post/173096796). seems inherently pleasurable/generative to be given flawed/incomplete machine-generated instructions & have to interpret/repair them while following along to make something
1
4
21
we also propose some new applications for existing technologies: for instance, leveraging predictive ML models (like those used in ’s level design support tools) to push users away from “obvious” or “stereotypical” design decisions
Quote Tweet
Short update on our AI Level Editor “Morai Maker”, showcasing work primarily by @thechillhammer to be presented at #EXAG19 this year. youtu.be/G1Md-wqh1vk
Show this thread
1
2
7
and we apply these design patterns to three case studies of systems developed here at UCSC: - Redactionist (formerly known as #blackoutgen), an erasure poetry creation tool - Germinate, an AI-supported game design sketchbook - Why Are We Like This?, a co-creative writing game
Screenshot of the Redactionist user interface, showing a partially constructed erasure poem: "an experimental poet means a nobody".
Screenshot of the Germinate user interface showing a procedurally generated game based on a user's explicitly declared design intent
Screenshot of a system diagram depicting how AI is used to provide the user with suggestions for how a written story could be developed or continued in Why Are We Like This?
1
5
18
in each case, the system was originally developed as a casual creator, but the addition of reflective creator design patterns helped us facilitate a more reflective creative experience & prompt users to make decisions they might not have made otherwise
1
2
6