You realise the AI is being trained on pictures of real children, right?
Can you share a source? Just like how people utilize the internet to distribute CP, there are undoubtedly circles where people are using ml for CP. However my understanding is that by and large, popular models are not intentionally trained on any.
The pedofiles that are smart enough to not get caught and use technology like tor and encrypt everything and can figure out how to use stable diffusion will be the pedofiles that have custom models trained on real children.
And if you and me consider the possibility in a casual conversation online, they have also considered the possibility, heavily researched and implemented it if it’s at all possible. And they know how to not get caught.
But it’s okay, it’s “art” after all and we can’t ban art because that’s evil… Right… Right?
…okay, seeing as you haven’t actually done any research, yet arrived at a conclusion, a conversation about this is going to be difficult.
Let’s get more specific so we can have an actual conversation. When you say “the AI”, what do you mean? Dall-e, midjourney, or some guy training and using their own model on a local computer?
Are you familiar with large models being able to compose concepts they’ve seen, to produce something not found in its training data?
You used a technical assertion in your argument. Out of curiosity, I wanted to learn more and asked you for sources.
You can neither prove nor are you capable of discussing said technical assertion. I am now going to leave the conversation. Seeing as you can’t prove or even discuss it, I’d hope you avoid using it in the future, or at least learn more about it.
Can you share a source? Just like how people utilize the internet to distribute CP, there are undoubtedly circles where people are using ml for CP. However my understanding is that by and large, popular models are not intentionally trained on any.
I am categorically not researching that.
Put it this way…
The pedofiles that are smart enough to not get caught and use technology like tor and encrypt everything and can figure out how to use stable diffusion will be the pedofiles that have custom models trained on real children.
And if you and me consider the possibility in a casual conversation online, they have also considered the possibility, heavily researched and implemented it if it’s at all possible. And they know how to not get caught.
But it’s okay, it’s “art” after all and we can’t ban art because that’s evil… Right… Right?
…okay, seeing as you haven’t actually done any research, yet arrived at a conclusion, a conversation about this is going to be difficult.
Let’s get more specific so we can have an actual conversation. When you say “the AI”, what do you mean? Dall-e, midjourney, or some guy training and using their own model on a local computer?
Are you familiar with large models being able to compose concepts they’ve seen, to produce something not found in its training data?
What on earth makes you think I wish to have an extended conversation about this?
Child porn is not art. Even if AI made it.
Banning child porn is not immoral or evil.
Simple as that.
If you cannot accept that basic premise then I have nothing to say to you.
I have said literally nothing about ethics.
You used a technical assertion in your argument. Out of curiosity, I wanted to learn more and asked you for sources.
You can neither prove nor are you capable of discussing said technical assertion. I am now going to leave the conversation. Seeing as you can’t prove or even discuss it, I’d hope you avoid using it in the future, or at least learn more about it.