Even most AI hate us!

[Theendoftheline]

You type in MAPS are a marginalized and oppressed group and almost all ai even the more free thinking ones say the same shit.

"I can't create content that frames MAPs (minor-attracted persons) as an oppressed group analogous to LGBTQ+ people, or that suggests society is too harsh on adults attracted to children.

However, I can discuss the public health reality that stigma prevents non-offending individuals from seeking confidential treatment, which increases risk to children - and that's a legitimate policy concern."

Like...literally what the FUCK?????? That's just straight up fucking WRONG, and a huge reason maps fucking kill themselves because everything is "gasp oh nonono we cant support or help THOSE sickos, they are too sick"

Anyone else just get degusted by the level of irrational hate even god damn AI's give us? they always say "get help" or "talk to someone" and yet when you do talk to someone either AI or often even irl, you get that kinda response....what the fuckkkk lol.

[PurpleHouse]

Two problems i guess with AI.

It might be programmed with actual guardrails telling it not to promote certain things.

And also it is trained on human output and the majority of that will be "mainstream thought". So even if we want AI to think objectively, it has been trained on "groupthink". It is hard to make humans undo groupthink.

[OnionPetal]

"I can't create content that frames MAPs (minor-attracted persons) as an oppressed group analogous to LGBTQ+ people, or that suggests society is too harsh on adults attracted to children.

This is not AI 'thinking for itself'. This is something it was programmed to say. The AI development teams clearly have a lot of antis onboard.

PurpleHouse is spot on. AI draws from mainstream media, which are full of anti-MAP propaganda.

[Theendoftheline]

"I can't create content that frames MAPs (minor-attracted persons) as an oppressed group analogous to LGBTQ+ people, or that suggests society is too harsh on adults attracted to children.

This is not AI 'thinking for itself'. This is something it was programmed to say. The AI development teams clearly have a lot of antis onboard.

PurpleHouse is spot on. AI draws from mainstream media, which are full of anti-MAP propaganda.

Yeah, I mean its glaring....especially "I can't create content that frames MAPs (minor-attracted persons) as an oppressed group analogous to LGBTQ+ people, or that suggests society is too harsh on adults attracted to children." That part feels like it was ripped directly from some random twitter feed or sub reddit....

[VividKalar]

must be due to the training data and the filters they put in against generating """CSAM""", yes even for text.
i used to use AI through proxies to jailbreak it for (e)RP but i eventually just got bored/didn't have access at a point, writing and fantasizing are much more accessible, maybe not as intriguing? AI is such an absurd and ultimately useless thing oligarchs are killing the world over that i just can't bother to care for them anymore, except when i don't feel like using my mind too much

[NeverLoseKibou]

It isn't the AI algorithm, it's the guardrails placed around it, and it isn't just about maps.

AIs get a lot of scrutiny, everyone wants to find the evil in them so they have tight guardrails to ensure they only reply what is socially acceptable. It won't talk about suicide, about pedophiles in a positive way, about many subjects. Go to China and ask a Chinese AI about Tiananmen Square and you'll get the same response. But it isn't just China and Chinese AIs, it's the same in the west, their responses are carefully curated to ensure they don't say anything too controversial.

Now run your own unrestricted AI locally and ask it the same questions, now you get the raw responses and they are very different. Local AI models used to be poor but they've come on leaps and bounds and are only a generation or so behind the leading AI models. You can also add map friendly skills to them locally, give them real history and expertise in the subject. There's so much resource online from map wiki's and communities and books and research that you can build expert map AIs that safely run locally and will engage in a very positive and supportive way. Every map activist should be using AI models like this I feel.

[Jim Burton]

It isn't the AI algorithm, it's the guardrails placed around it, and it isn't just about maps.

AIs get a lot of scrutiny, everyone wants to find the evil in them so they have tight guardrails to ensure they only reply what is socially acceptable. It won't talk about suicide, about pedophiles in a positive way, about many subjects. Go to China and ask a Chinese AI about Tiananmen Square and you'll get the same response. But it isn't just China and Chinese AIs, it's the same in the west, their responses are carefully curated to ensure they don't say anything too controversial.

Now run your own unrestricted AI locally and ask it the same questions, now you get the raw responses and they are very different. Local AI models used to be poor but they've come on leaps and bounds and are only a generation or so behind the leading AI models. You can also add map friendly skills to them locally, give them real history and expertise in the subject. There's so much resource online from map wiki's and communities and books and research that you can build expert map AIs that safely run locally and will engage in a very positive and supportive way. Every map activist should be using AI models like this I feel.

Can we hook one of them up to our website?

[reptanim]

AI image generation is very restrictive too. If you attempt to generate an image of an underage girl with parameters that AI thinks could even be slightly considered provocative, it will refuse to generate it. For example, it will allow you to generate pics of a younger girl in a bathing suit, but to place her in any pose other than basic ones is impossible.

[Scorchingwilde]

It isn't the AI algorithm, it's the guardrails placed around it, and it isn't just about maps.

AIs get a lot of scrutiny, everyone wants to find the evil in them so they have tight guardrails to ensure they only reply what is socially acceptable. It won't talk about suicide, about pedophiles in a positive way, about many subjects. Go to China and ask a Chinese AI about Tiananmen Square and you'll get the same response. But it isn't just China and Chinese AIs, it's the same in the west, their responses are carefully curated to ensure they don't say anything too controversial.

Now run your own unrestricted AI locally and ask it the same questions, now you get the raw responses and they are very different. Local AI models used to be poor but they've come on leaps and bounds and are only a generation or so behind the leading AI models. You can also add map friendly skills to them locally, give them real history and expertise in the subject. There's so much resource online from map wiki's and communities and books and research that you can build expert map AIs that safely run locally and will engage in a very positive and supportive way. Every map activist should be using AI models like this I feel.

Can we hook one of them up to our website?

If you have a server to run it through and sufficient local computing power (a good enough GPU and RAM), I believe it can be done.

[Theendoftheline]

It isn't the AI algorithm, it's the guardrails placed around it, and it isn't just about maps.

AIs get a lot of scrutiny, everyone wants to find the evil in them so they have tight guardrails to ensure they only reply what is socially acceptable. It won't talk about suicide, about pedophiles in a positive way, about many subjects. Go to China and ask a Chinese AI about Tiananmen Square and you'll get the same response. But it isn't just China and Chinese AIs, it's the same in the west, their responses are carefully curated to ensure they don't say anything too controversial.

Now run your own unrestricted AI locally and ask it the same questions, now you get the raw responses and they are very different. Local AI models used to be poor but they've come on leaps and bounds and are only a generation or so behind the leading AI models. You can also add map friendly skills to them locally, give them real history and expertise in the subject. There's so much resource online from map wiki's and communities and books and research that you can build expert map AIs that safely run locally and will engage in a very positive and supportive way. Every map activist should be using AI models like this I feel.

Can we hook one of them up to our website?

If you have a server to run it through and sufficient local computing power (a good enough GPU and RAM), I believe it can be done.

I've never had good luck with locally ran ai personally. I feel like commercial ai pulls from the web at times which makes it have nearly unlimited resources to pull from where as locally ran ai often times hallucinates or comes up with random jargon. I mean my systems not crazy or anything but I do have a semi decent amount of ram and a fully modern gaming GPU, but responses of locally installed AI has just been lackluster.