GMs, how have you used ChatGPT to help run your games?
GMs, how have you used ChatGPT to help run your games?
ChatGPT has been a lifeline for me as a GM with little spare time to prep and far too grand ambitions for the scale and scope of (D&D) campaign I want to run. I'm curious how other GMs have found ChatGPT and similar AI tools useful or helpful in running their own games. I'll share my own workflow below as a comment, and I hope others find it useful. I'm especially interested in any ChatGPT prompts you have found worthwhile, and you can see some of my own prompts in the examples I'll share shortly.
You're viewing a single thread.
I'm glad this kind of tooling works for people, but the idea of it is very alien to me. I've also never been fond of random tables. Somehow it feels "wrong" to use a table or chatgpt in a way that stealing ideas and mashing them into "what if the head vampire is basically nick fury but he has to take care of a baby-yoda like thing like from the mandalorian" doesn't.
I can understand the skepticism, but if helps things up at all, I would clarify that I think exceedingly few people who use random tables (even all the way back to the 70s) really use them as a final say in what's happening next. Most who use it probably appreciate it more as a way of dislodging tired tropes and tricks they find themselves always falling back on.
I think of it as one more source of inspiration to keep in your toolbox, alongside the evergreen "steal from media you like." That said, ymmv as with all things in life, and the best prep is the one that works for you!
See, in my campaign I do what you're talking about with regards to mashing ideas together. I just use ChatGPT to get a detailed description of a room/character and to help me to plan out how I ought to glue the ideas together in a way that fits my setting/theme. Sometimes it even adds additional influences to the mix to make the idea seem more embedded in the setting and original.
If you just use it on the prep side, you maintain full editorial control, and you can reply to its bad ideas with "let's do it more like this:" and it'll revise its descriptions/ideas to match.