How do we AI-proof online boardgames?

i think over time it'll organically gatekeep itself since AIfags are consumerist bugmen who need to chase a quick high and a shot of "content", they'll get bored and move on to the next thing
they can't commit to anything or else they'd actually foster a skill instead of faking it

>>715058329 (OP)
0%

>>715058329 (OP)
Why would there be a need? Board games (online games as a whole, really) are only fun if you're playing with people you know because people take things too seriously these days; it's all about winning rather than having fun. Card games are the worst of the lot because you have people netdecking. So find a good group of people to play with and there's no need to AI-proof online games. Heck, the tryhards basically function as AI. Can you really tell the difference between a person using the highest win rate things and an AI using the same strategy?

>>715058668
>Can you really tell the difference between a person using the highest win rate things and an AI using the same strategy?
The difference between a nyagger and an AI is astronomical. Only a nyagger can play with soul (i.e. furiten 90% of the time).

Play games that don't have perfect information and even an AI model won't be able to fully plot out optimal moves because of a little thing called bluffing. At that point all they can do is play the odds.

>>715058775
I guess in gambling that would be true.

>>715058329 (OP)
That's like asking how you make chess engine-proof. You don't. You have to be selective who you play with and ban people whose inputs are suspiciously machine-like. Just like with chess, if you play with randoms then you must accept you're going to sometimes be playing against losers outsourcing their thinking to bots.

>troonjong
>boardgame
begone weebshit

>>715058668
>le play for fun
I have fun by winning

>>715058329 (OP)
Turn timer that's only a few seconds. It'll take too long for people to tab out/get their phone, ask AI what to do, gen the response, and act. Only intuitive human reaction will be possible, at least until vision models advance enough that you can just screen-share with your LLM and it gives live commentary.

Just play through Tabletop Sim

>>715058826
If you're playing online with randoms the experience is just as bad. The likes of Town of Salem has a meta game to follow that makes the entire thing feel both stifling and samey, which kinda defeats the purpose of the game. You can totally teach an AI to follow the meta gameplan in those games and probably have better results than a player, especially if they're able to accurately able to keep track of information being presented.

>>715059108
So does the AI. You're literally a bot bro

>>715059113
>at least until vision models advance enough that you can just screen-share with your LLM and it gives live commentary.
anon.....

>>715059240
If a game is simple enough that complex interlocking parts have gotten distilled into a flowchart to follow, there's not that much more to do in the game.
Most social deduction games are simply logic deduction where you can only partially obscure the facts.

>>715059452
This. At worst you can solve them by following a meta

>>715059452
That's what online games turn into, which is why online games are so unfun these days. Either a game gets solved or as close it it as possible by removuing the deduction from a social deduction game. Try playing one of those games and not following the meta as see how quickly you get off'd because you're "sus".

>>715059693
Theres a reason you are meant to play party games with friends.

Ameritrash games with lots of dice rolling are AI proof