I’m thinking of incorperating a phrase in tools and potential future games I write that indicate that no AI was used in the writing process (not excluding secondary tools, like the compiler).
Do you mean “not including secondary tools like the compiler”? That part may be harder to track—I can currently say with some confidence that the Dialog compiler has no AI-generated code in it, because the only changes to the compiler inner workings have come from me and Susan Davis, and I believe neither of us uses LLMs to code. And I can say with some confidence that the ZIL compiler does, because Tara McGrew has been very open about her LLM use. But it’s not hard to imagine an alternative world where other people submit LLM-generated pull requests to Dialog and Tara doesn’t discuss LLM use openly on this forum.
Regardless, I wouldn’t mind a label like this, especially if someone artistically inclined were to design something cool for it. I’m not a purist about avoiding games with LLM-generated code, but I don’t intend to use it myself, so if other people do care then I won’t pass up the opportunity to pat myself on the back a bit.
The question of what human made means reminds me of a puzzle maker who uses hand and power tools to shape masters either from scratch or by modifying parts from mass produced puzzles and then making molds and casting resin copies of those masters as needed claiming designers who use CAD software to design puzzles and 3-D print those designs didn’t actually make the resulting puzzle.
If we are talking about interactive fiction stories/games, what I care about the most is that the narrative content I am reading is written by a human being.
If the compiler and/or programming language under which a human made story has been developed was co-authored by an LLM… so what?
I feel like if you want the whole toolchain to be AI-free, It’s much more about you making a statement and expressing a strong opinion about AI in general. Which is fine, everyone has their rights to their views and beliefs, but I feel this goes beyond what such a “seal” should be about.
LLMs, in particular Anthropic’s Mythos model family, is currently significantly revolutionizing the life sciences industry, from the discovery of new medicine to cancer research and rare diseases. I doubt anyone would reject a cure for cancer just because it was co-developed with an LLM.
Henrik’s stated goals seem pretty unambiguous. Perhaps it would be appropriate to create a new thread about the revolutionary properties of LLM, if that’s a desirable discussion to have. If we don’t have such a thread already, that is. I have the AI tag muted.