AI rule for Spring Thing: How to make a rule that is enforceable and fair?

Competitions Spring Thing
41

I’m sorry, because this is the first time I’m weighing in on this thread and I fear I will come off as rather hostile but if you cannot be bothered to write your own argument against the banning of AI, then why on earth should I care when any argument I have for the banning of AI has had to come from my own thoughts and opinions? Are we outsourcing emotion and reasoning and logic to the machines now?

My thoughts on this subject are that all AI, including AI assisted coding, should be banned. I would rather see a beginner’s clumsy art, grammar-error riddled sentences, code that’s redundant and half-broken and buggy – I would rather see something that is utterly human in its seeking to be than something that is perfect and utterly devoid of life.

“But what about a disparity of resources? Some people aren’t good artists, some people can’t write and still want to participate and…” learn. Try and fail and learn. Writing, art, code – all of these are skills that you must use to get better at them. You do not get better at them overnight or in some single stroke of genius. You will be no better for using AI to write your game, you will learn nothing, you will not get better, you will stagnate and do nothing but feed a machine your ideas forever.

20 Likes
42

That’s true but I don’t see what relevance this has to what anyone should do in a hobbyist space. Companies do a lot of things in pursuit of profit that simply don’t make sense for hobbyists to emulate because we don’t have the same goals.

17 Likes
43

But I do think that we should not shy away from a tool, just because it has the tendency to produce garbage if not used wisely. I think the ‘company’ statement was really to highlight that this tool is ‘out there’ and is generally available to all of us - people can make their own decision as the whether they want to use it or not . Being one of the older generation, brought up in the 8 bit era, I am not adverse to trying new technology, especially if it assists me in my own goals - I, for one will continue to use it; most of my games never get to the ‘public’ arena anyway, so I am not too bothered if it would be prevented in entering a comp. If I was so inclined, then I would state exactly how I used AI and let the comp ‘adjudicators’ decide if it was acceptable or not. It does open up a massive ‘can of worms’ anyway (as one of the ‘imps’ put it ‘Can of worms : opened’ :))

1 Like
44

I’m sorry @30x30 I was not using co-pilot to write a response for me, the ‘exercise’ was more to see what ‘argument’ AI produced against the use of it in IF games. As @mathbrush says, I am not sure it was a ‘valid’ contribution to the argument - although, I do agree with some of the statements (now that I have actually read it - @mathbrush was correct, it was copied and pasted with no inspection first!) .

45

My dude.

23 Likes
46

:neutral_face: :backhand_index_pointing_up:

edit: actually, it did worse than that, it distracted from the thread’s discussion and made real humans have to read and address the words an unthinking machine that put no effort into the creation of their response. it contributed a net negative to the thread. congratulations.

16 Likes
47

la commedia è finita!

5 Likes
48

Posting AI generated content like this is against intfic’s code of conduct.

EDIT: Which is here: FAQ - The Interactive Fiction Community Forum

9 Likes
49

I tend to agree (italics mine).

:+1::+1::+1:

1 Like
50

Even if there are problems detecting AI generated code, allowing it is potentially green-lighting or endorsing its use.

If you don’t want people to submit AI generated code then I think it should still be valid as a rule. The majority of people will probably follow the rules, and the submission process itself can probably be used to filter out the majority of low effort entries.

2 Likes
51

I’m not clear what the argument is for banning its use.

1 Like
52

@mathbrush
I watched an interview with a teacher recently who had “the gall” to suggest in-class essay writing. :wink:

Such a simple solution. But alas, I doubt many teachers today even know that’s a thing you can do.

It took decades for them to get the balls to start banning cell phones in classrooms. Let’s see how long it takes for them to outsmart inexperienced children.

(Hopefully, it happens before AI replaces their jobs.)

3 Likes
53

I feel I must reply to this, and to profusely apologise if I have offended or upset anyone by posting the AI response to the objection of using AI in IF games. I did this as an exercise to see how, and with what arguments, it would ‘defend’ the use of itself. It was not meant as hateful at all. I do feel however, that some of the point it made are valid - that it is a creative tool and should be used in a way to help ‘create’ : It is not meant to replace either artist or authors and it never will - art comes from human beings not machines.

1 Like
54

Thousands of words in this thread, not sure I caught it all.

IMO, it’s too easy to vote for an anti-AI rule with no clear enforcement mechanism. “Someone should bell the cat!” But, who, and how?

Brian, are you going to act as the AI police, investigating games where AI is suspected? I wouldn’t sign up for that, if I were you.

I think IFComp took the right approach with requiring authors to disclose AI usage, with approximately no enforcement mechanism for lying.

I would add a warning that “games that are substantially generated by AI are very unlikely to win.”

Voters can then incorporate the AI disclosure into their votes, can vote against games that are bad (regardless of whether they’re AI).

Liars will be rare, and they’ll get caught by making a bad game, not by requiring someone to do a forensic investigation.

4 Likes
55

Is there an airtight enforcement mechanism for any of the other Spring Thing rules, though? It seems to come down to “most people entering the competition will follow the rules just because they exist, regardless of enforcement, and the benevolent dictator can remove any egregiously bad actors”, and that seems to work pretty well.

17 Likes
56

It is too difficult to split the prize money 10 million ways.

2 Likes
58

The organizer can straightforwardly detect whether a game is new finished IF. I suppose there is an enforcement challenge with ensuring that you didn’t plagiarize the game, but plagiarism requires evidence of another work you’ve plagiarized; it’s easy to verify once the proof is exhibited.

AI is not detected that way. It’s detected entirely by feel. Does it feel like AI slop? Does it make excessive use of emdashes or “delve”?

That means it’s inherently costly to enforce an AI rule, and the vibes on enforcing the rule are totally off.

Do you enjoy playing AI police? Is that the position you want to take relative to participants? AI policing requires an attitude of mistrust, suspicion. Assume they’re trying to pull the wool over you, and then find the evidence.

Enforcing that rule sucks. You will make mistakes, guaranteed, both ways. You will have to fight with liars who are obviously lying. To whose benefit?

1 Like
59

That’s an interesting point.

Right now there are two comparable scenarios.

One, for Spring Thing, is the pre-existing release rule. Pre-existing games are not allowed, and this year there were three games that fell under that rule. I contacted them dashed them about it. One of them was able to remove the game from the public before it got a lot of attention and were able to enter. Two were not, and it was a bit stressful, but I didn’t include them in spring thing (but did mention them in the main post).

I also moderate the interactive fiction subreddit where Ai-related discussion is banned (and forwarded to Chatgptgaming). There’s a script on the page to prevent any topic with AI, GPT or LLM in the title. Several people post anyway, and the posts get removed because they don’t hide it. They acknowledge ai use openly. Some slide through because they only mention ai on the pages they link to, but they get reported and blocked.

But some posts live on the edge. There’s one guy that posts “how to make a game” videos every single day that has ai-made thumbnails. I don’t remove them because the actual content isn’t AI and I feel bad that his videos have low views.

So imagine this would be similar. There would be some entrants each year who cite AI in the credits or have their code on a public github that uses AI or have a Twitter account that mentions making a game with AI. In those situations, disqualification would be clear. In situations where I couldn’t prove it, I could ask them and if they deny it, I would likely just have to let them enter. So yes, this is hard to enforce and is just meant to discourage.

That said, I’ve noticed that a lot of people that use AI really overestimate how much other people like and respect AI, and underestimate how obvious AI usage is.

People who are good at hiding ai usage aren’t the target of this policy; this is just for the lowest-hanging fruit.

I do agree that arguing about it is fruitless. I publicly accused one comp participant years ago of ai usage and was told that that was inappropriate (rightly so). I messaged the author and they admitted using ai to regurgitate two game books they had fed into it with slightly different wording, and they withdrew their game from IFcomp out of fear of being caught and getting in trouble at their work. I wouldn’t want something like that to happen again.

16 Likes
60

The rule doesn’t need to be enforced per se. If AI is unwanted in the Spring Thing (and mathbrush decides that), then a rule can say that it’s not allowed, and act as a dererrent. It doesn’t automatically mean that he has to look closely at each game to see if they’re made with AI, but if it’s obvious, then they can be made an example of and be banned. If a game made with AI is submitted regardless, and it’s not obviously AI, and nobody notices, then whatever, right? What’s the desired spirit of the law here?

In my country it’s technically illegal to drink in public spaces. Everybody drinks in parks during summer, and nobody cares, the police least of all. But if someone makes a ruckus and are a public nuisance, then the public drinking law can be used as a pretense for breaking up the party.

I’ve never made an IF myself, by the way, and as such these rules aren’t directed at me. But I am completely uninterested in playing a game that’s made by AI (just like I’m uninterested in reading a book written by AI, or listening to music by an AI, etc). But if someone manages to make a game with AI as a helping tool, in such a way that I don’t notice that it’s AI, then I’d probably play it. Maybe I’d even have been “tricked” into enjoying it. No harm, no foul, I guess. But if I knew it was made with AI to begin with, I’d probably not have even bothered to play it, because I’d rather use my limited time to play a game made by a human. Not sure what that says about enforcement (the dishonest author might win the thing with a rule-breaking game, and if they were honest I’d be deprived of a good experience!) but I’m not sure how much that matters. “Don’t ask, don’t tell”?

Edit: Mathbrush posted a similar argument while I was writing!

9 Likes
61

i think we’ve lost sight of the original post, which asked something like “what should the rules be regarding use of AI?”

there was nothing said about enforcement or how difficult it is. to my eye this is no different than making a rule about plagiarism, or using someone else’s IP, or making this year’s comp theme “cowboys and time travel”.

i see no problem with making rules against using AI in (non-trivial) content generation or coding. and i also see no problem in this policy evolving over time as the tools themselves evolve.

7 Likes