THIS Is Why We Should Keep AI Out of D&D

[GrandMajora]

Because teaching AI how to play the role of villain can go very, very wrong, very, very fast.

https://www.youtube.com/watch?v=0GCVTzNOaNg

[Heu, Iterum Id Feci]

You mean if you ask an LLM to act like an evil AI this LLM will then act like an Evil AI from popculture and texts? Color me entirely not surprised...
Also ChatGPT isn't an AI in the same sense Skynet or Matrix is...and even in both those cases, while they seem Evil and are portrayed as the antagonist, they are not Evil in exactly the same way a human would be evil.
it's a fundamental misunderstanding how ChatGPT works...It's not very intelligent as in making completely new decisions. It's a prediction model in a sense. It tries to predict what a human might say given the prompts and the trainingdata it has.
I'm also tired of this "no AI" attitude. AI is a tool. A very powerful one for those, that know how to use it. And like any tool it can be used for good and for evil.
Sure, giving the AI the full control over your campaign will most likely end in a very predictable and boring campaign, lacking interesting story twists (well not accounting for hallucinations or the AI simply loosing track of itself) etc. and as such using only AI should be avoided.
However, LLMs such as ChatGPT can be used to help you in campaign prep, if you know how to use it.

[GrandMajora]

Originally posted by 1337_h4x0r_xXx_deathlord666_xXx:

You mean if you ask an LLM to act like an evil AI this LLM will then act like an Evil AI from popculture and texts? Color me entirely not surprised...
Also ChatGPT isn't an AI in the same sense Skynet or Matrix is...and even in both those cases, while they seem Evil and are portrayed as the antagonist, they are not Evil in exactly the same way a human would be evil.
it's a fundamental misunderstanding how ChatGPT works...It's not very intelligent as in making completely new decisions. It's a prediction model in a sense. It tries to predict what a human might say given the prompts and the trainingdata it has.
I'm also tired of this "no AI" attitude. AI is a tool. A very powerful one for those, that know how to use it. And like any tool it can be used for good and for evil.
Sure, giving the AI the full control over your campaign will most likely end in a very predictable and boring campaign, lacking interesting story twists (well not accounting for hallucinations or the AI simply loosing track of itself) etc. and as such using only AI should be avoided.
However, LLMs such as ChatGPT can be used to help you in campaign prep, if you know how to use it.

I think you might have missed the joke entirely.

And while I understand that AI is ultimately just a tool, it is unfortunate that there are so many people out there who are too lazy to use it properly. Like, people say they are using AI in art to serve as 'references,' but they don't even put in the bare minimum effort of fixing the warped distortions and model inconsistencies that AI programs generate. They just go with whatever the AI produces, no matter how off putting or unnatural it seems.

Then there's the problem with AI Music, which all seem to have this electronic voice filter over the vocals, that has the irritating side effect of making every song sound like it's being sung by the same band!

Like, seriously, stop making AI music! I genuinely can't tell the difference between one vocalist and another, because they all sound identical!

[アンジェル]

Originally posted by GrandMajora:

Originally posted by 1337_h4x0r_xXx_deathlord666_xXx:

You mean if you ask an LLM to act like an evil AI this LLM will then act like an Evil AI from popculture and texts? Color me entirely not surprised...
Also ChatGPT isn't an AI in the same sense Skynet or Matrix is...and even in both those cases, while they seem Evil and are portrayed as the antagonist, they are not Evil in exactly the same way a human would be evil.
it's a fundamental misunderstanding how ChatGPT works...It's not very intelligent as in making completely new decisions. It's a prediction model in a sense. It tries to predict what a human might say given the prompts and the trainingdata it has.
I'm also tired of this "no AI" attitude. AI is a tool. A very powerful one for those, that know how to use it. And like any tool it can be used for good and for evil.
Sure, giving the AI the full control over your campaign will most likely end in a very predictable and boring campaign, lacking interesting story twists (well not accounting for hallucinations or the AI simply loosing track of itself) etc. and as such using only AI should be avoided.
However, LLMs such as ChatGPT can be used to help you in campaign prep, if you know how to use it.

I think you might have missed the joke entirely.

And while I understand that AI is ultimately just a tool, it is unfortunate that there are so many people out there who are too lazy to use it properly. Like, people say they are using AI in art to serve as 'references,' but they don't even put in the bare minimum effort of fixing the warped distortions and model inconsistencies that AI programs generate. They just go with whatever the AI produces, no matter how off putting or unnatural it seems.

Then there's the problem with AI Music, which all seem to have this electronic voice filter over the vocals, that has the irritating side effect of making every song sound like it's being sung by the same band!

Like, seriously, stop making AI music! I genuinely can't tell the difference between one vocalist and another, because they all sound identical!

That is more a personal issue of yours, if you cannot distinguish the voices of the Hatsune Miku cast, though. They even have different personalities.

[GrandMajora]

Originally posted by アンジェル:

That is more a personal issue of yours, if you cannot distinguish the voices of the Hatsune Miku cast, though. They even have different personalities.

Hatsune was created long before this AI generated content got so popular. She's a high quality product.

What I'm talking about is the deluge of amateur work that's taken over Youtube lately.

[Heu, Iterum Id Feci]

It'll sort itself out over long or short. Those that don't put in the effort won't have longterm success. Sure, they might upset the balance a bit in the short term and we are in the midle of that phase, especially as AI tools are still under high fluctuation, but long term, those that can make high quality products wether be it with or without AI will prevail (though those using AI will have a major advantage on their side).
Those that can't deliver a minimum quality, no matter which tools are used, are gonna be relegated to the bargain bin...

[Pan Darius Cassandra]

Wow, that's amazing.

Soon, AI won't just be playing games - it will be creating them from prompts.

https://www.worldlabs.ai/

[God Queen 158]

Why would you add AI cringe grimdark dialogue when we already have the Dark Urge cringe dialogue.

[Pan Darius Cassandra]

Originally posted by God Queen 158:

Why would you add AI cringe grimdark dialogue when we already have the Dark Urge cringe dialogue.

It's hilarious.

[GrandMajora]

Originally posted by Pan Darius Cassandra:

It's hilarious.

Agreed.

[Driss]

How did it go wrong? It seemed to behave quite well in its role.

[Jean-Maurice Nya]

I like leaving SUNO's AI work untouched. It feels well. It feels like it is a real R&N artist at work.