Well, I’m not sure of alot of things, but I’m fairly certain that response was not written with an LLM. That was impressive. Sincerely.
10 Likes
I don’t see how the protected characteristics of an author, being used to discriminate against their work, compare to using AI to create work. The point of these competitions is to test and challenge the human creativity and ability to come up with great storytelling and great writing. Women, LGBTQ, ethnic minority and other authors are all human beings with the creativity and ability mentioned above. AI (or more specifically machine learning) bypasses that completely, the process is completely programmatic and therefore precognitive (I think that’s the right word?). It’s basically equivalent to having someone use stockfish and other chess algorithms when participating in the international chess championships.
13 Likes
I’m going to print this out and frame it.
7 Likes
I love how people here are respectfully hating each other’s guts. In other forums, this would have turned into a bare-knuckle brawl, but not here… we’re all civilized an’ shit. Way to go, us! 
11 Likes
It’s ‘cos’ @HanonO keeps us in check! I can feel his eagle eyes watching over this! 
9 Likes
That’s a valid opinion.
Another perfectly valid opinion is that using AI to help generate ideas helps creative people be more creative. And AI is improving. At a rapidly accelerating rate. There will be a point at some unknown time in the future where AI can match or surpass human creativity.
While there certainly are people who do not want to see AI content, there are also people who do not mind it or even are curious about it and seek it out.
As far as IFComp goes, I hope it remains neutral regarding generative AI, but even if it outright bans that type of content, I believe there is a place in the world of IF for generative AI, even if it is not everyone’s bag of tea.
-v
2 Likes
You need to write this interactively as your submission for IFComp 2025.
1 Like
The biggest difference here seems to be whether one views that with anticipation or mourning.
5 Likes
One idea might be to do an AI comp. There are several people on the forum that are interested in AI, but feel censured in the current competitions. An AI comp would attract only judges interested in AI specifically.
I’m not suggesting this is in a ‘go make your own comp and stay away’ kind of way, but more in a ‘parsercomp’ way. Parsercomp never detracted from IFComp, but it was a way to boost and honor a particular kind of game. An competition of games made with AI (and asking only for interested judges) could be a way for people to show off what AI can do without getting attacked by others.
I myself am AI-adverse and would not enter a game in the competition (as my primary reason for writing is to develop my own skills until I can prove to myself that I like my own writing and am proud of it even more than achieving public recognition, which is my second goal), but it could be worthwhile.
12 Likes
This isn’t an argument for or against any particular rule, but I think it’s the prerogative of any group or competition to define what they want their group to be, even if it’s entirely voluntary or even unenforceable. IFComp’s new disclosure rules seem to me to be trying to be inclusive of as many people as possible and to not take sides, which is totally understandable.
I should also admit that my original post was at least partly the predictable response to the statement to the effect of “feedback can be made after the competition”. I will complain wherever and whenever I want to thank you very much 
This whole conversation reminded me of the Colorado State Fair controversy from 2 years ago, when an AI generated painting won (and which had disclosed the use of AI). Perhaps the lesson is that disclosure is not going to prevent controversy, but maybe a NYT article will be forthcoming.
2 Likes
Honestly, this is probably the most constructive comment made so far.
6 Likes
I completely agree with everything you said.
I was just pointing out (in my pedanticaly logical and apparently tonedeaf way) that having a problem with particular content is a different thing than having an issue with a particular origin of content. I don’t have an issue with AI being flagged in competitions, I just found the analogies to existing situations somewhat strained.
Agreed. The comparison to chess competition is an apt one I think.
3 Likes
Anticipation and mourning are two possible reactions, out of many. Another reaction is for people to attempt to wield political power or social influence in a mostly pointless attempt to prevent it, kind of like they are doing with cloning and a million other things that are new and frightening. And yet another is just to observe and see what happens, neither sad nor greatly impressed by anything by AI so far, but also, not trying to stop it.
-v
2 Likes
If you would like a better comparison, the most apt one I can come up with is GMO labelling. This is another case where you have a controversial production process (also frequently misunderstood by both detractors and advocates) where disclosure has been a tenable solution. People can make their own choices to avoid it or not, and whether those choices are based in logic or not (ex: not wanting to support Monsanto vs thinking GMOs will give you cancer) isn’t anyone else’s concern.
In the future, I might politely suggest refraining from making comparisons to real-life bigotry in online arguments except in exceptional cases. It rarely serves the intended purpose (unless the intended purpose is to rile people up).
9 Likes
I’d like to think whether something succeeds or not is not the only measure of whether it was pointless. By that measure, every human endeavor may ultimately and eventually be pointless. Then again, what do I know.
Regardless, can we just go with Mathbrush’s idea and call it a day for now? I’m sure someone will start a new thread next week and we’ll have a chance to rinse and repeat then.
5 Likes
That is a far better example. I wish I had thought of it.
Bigotry in any form is terrible. It’s clear to me that I should’ve used literally any other example.
10 Likes
The main issue here is that the comment equated (cis white straight male authors vs women, LGBTQ, ethnic minority authors) to (human authors vs AI authors), which suggested that the differences between the 2 comparisons are identical.
This is definitely not the case as the two in the first comparison will both use the same creative process (brain power) to create works of interactive fiction and therefore the comparison here is zero (no difference between these two), while the two in the second comparison will not use the same process (humans use brain power, AI uses algorithms) and therefore the comparison here is anything but zero (major difference between these two).
To equate both comparisons to zero suggests that humans and AI are identical, which is not true and to equate both comparisons to non zero suggests differences between authors based on protected characteristics, which pushes an argument of bigotry based on humans and AI being different and that is what most likely caused offence.
However, it seems like you understood your error here so I’d chalk it down to a learning experience to not use an analogy of that nature again. Also, general advice, when making an analogical argument, make sure the analogy uses the same nature as the original point (ie. human vs computers in this case), like the chess argument I gave.
It’d be more like robot wars than IFComp but that could well be a good idea. It’d definitely be interesting to see how it goes. There’s a chess equivalent here in the form of the computer chess tournaments.
Agreed! 
5 Likes
Not much of a novel idea looking at past threads:
though none have bore fruits so far.
7 Likes
Part of the problem here is that “AI” is a hot buzzword right now and is getting slapped on everything, regardless of the actual type of algorithm. Of these, generative AI is the one that people have ethical concerns about.
I agree, and that was my first reaction to the question.
Upon more consideration however, I think a rule that specifically singles out generative AI and/or LLMs (large language models) still has fuzzy boundaries. I’m not necessarily against such a rule, but I do think it’s worth recognizing how difficult it is to draw the lines.
Virtually all spelling, grammar, and style checkers use statistical data computed from a corpus of text that was probably collected without anyone’s permission. (The lexicon used by the earliest Unix spelling checkers was supposedly generated from years of corporate email and cross checked with a probably-not-authorized electronic copy of a published dictionary.)
Almost all of these checkers now offer suggested edits. Many of those use some form of generative AI (admittedly, a more primitive form than the likes of ChatGPT). If I check the text my game produces with one of these tools and accept any of its suggestions–verbatim or with tweaks–then, to some degree, I’ve incorporated some AI-generated text into my game. Should that trigger a rule or not?
Statistical language translation is a close cousin to today’s LLMs. If I have an NPC that speaks in a foreign language, and I use translation software to craft their utterances, am I using generative AI in my work?
If I have a long brainstorming session with ChatGPT to come up with ideas for a plot or setting and then I author a story without any other AI assistance of any kind, do I have to acknowledge ChatGPT’s inspiration?
Then there’s the programming side of things.
Perhaps I’m porting my own Python code into Inform 7, and I get stuck on how to code a door that can be locked only from one side and unlocked only from the other. Is asking Copilot or ChatGPT is off the table? (Assuming, of course, that the AI knows enough about Inform 7 to help.)
Maybe I do a web search and get a link to a page where someone explains how to solve my problem. (Let’s ignore the question of whether the search engine used AI to direct me to that particular page.) There’s a decent (and growing) possibility that page was written by generative AI rather than a human blogger. Is it a problem to use the recommended approach? I think we’re already at the inflection point where people will often not know for sure whether they’re consuming generated content.
I don’t have enough experience with generated imagery to have an opinion about that. (And I don’t care for imagery in my IF, so I’m not that interested.) I suspect there are (for the moment) fundamental differences that make the AI-generated imagery slope less slippery than the text and code slopes. My gut says it would be easier to craft a rule that targets AI-generated imagery rather than a more general one that encompasses all AI-generated content.
4 Likes
In that case I’d say no: you’re doing research and discussing your plot with a friend (in this case an AI) to brainstorm, the same as people who make mood-boards of images that aren’t used in the final product for inspiration or go down a Wikipedia spiral for ideas and research. There has been plot generation/inspiration software before - not AI but more of a mad-libs database which can throw a monkey wrench at a situation like improv theater or help structure character motivations and goals.
If you include a direct quote from a blog or website or other work verbatim, though, you cite.
Where it gets fuzzy I think is when an AI or LLM comes up with actual player-facing prose that is used verbatim copy/paste in the finished product.
While that may seem to be the “torches and pitchforks” trigger for many people, my devil’s advocate impulse can see a situation where a game includes an actual AI character and uses GPT or something similar to improvise that dialogue for them specifically to get that uncanny AI vibe. In that case the AI likely should be disclosed/credited/cited as contribution to the writing - the same as if you tasked another person to write one specific character’s dialogue to make it tonally different from what you would write - you’d credit that contribution.
7 Likes