Hello Choice of Games forum! It’s been too long.
I was wondering if this discerning choice-based fiction audience could help with a science (?) experiment. The deadline for the AI in Interactive Digital Entertainment conference is fast approaching (abstracts due Friday), and so I whipped up 3 conditions for automatically generating choice-based IF through different prompts to ChatGPT. (One of these uses a tiny bit of Real AI itself, and the other two are basically controls, but still use ChatGPT.) Each code snippet generates a narrative that is exactly 5 choices long.
Choose what the story’s about with freeform text at the top, hit Start, and then make choices with radio buttons and hit Go.
Once you go through a story, please fill out the corresponding survey form, because I can’t get into AIIDE by just entertaining people. It might be art (?) but it’s not publishable without statistics! These links are also given at the end of each story.
Story A Survey
Story B Survey
Story C Survey
It would be great if I could get as many people as possible doing all 3 conditions. 5 choices! That doesn’t even get you through some dream sequences in Choice of Robots. Each survey’s pretty short, too, I think.
You can see the code at the site I used to host, but please don’t look before doing the surveys to avoid biasing yourself, and please don’t comment in this forum about what you find there. I promise to come back and explain what I’m trying to do at some future point once the survey is closed.
I hope nobody is upset at the prospect of AIs writing choice-based stuff - you’ll see from the experiment that the CoG stable of writers doesn’t have much to fear at the moment. It’s actually difficult to even get ChatGPT to consistently follow the rules of choice-based fiction. If it blurts unselectable choices at you, tells you what each choice is going to do before you pick it, or tells you to literally turn to a page, just roll with it!
Edit: Please consider starting with condition B or C! I have twice as many A responses as C responses, and I’d ideally like the counts to be balanced. (Of course, it’s ideal if you do more than one…) The overall entertainingness of your story probably depends more on your premise than the experimental condition. (They seem to do best with specific, non-bizarre, genre-related prompts that suggest a problem, like “a robot maker who has trouble relating to other people.”)
Edit2: The games and survey have been taken down - experiment’s over. Thanks to everyone who participated! An explanation of what this was is further down the comment thread.