undefined | Better HN

0 pointschabes12d ago0 comments

That consent of portrayed parties is impossible.

What is the solution there?

0 comments

Shouldn’t it be possible for AI to filter out that a request is made to portray a real person? That seems almost like a trivial task for a good model. I am sure every now and then something will slip through, but I bet one could make it very close to 100% effective.

nitwit00512d ago

Consider the difference between "Generate an image of Emma Watson", "Generate an image of Hermione", and "Generate an image of a female hogwarts witch and student". We're getting less and less specific, but those are all likely to get you an image of Emma Watson.

Your filter has to pick out that, while they did not ask for a specific person, the practical result is likely to be the same. That's going to be tough to get near perfect.

TheOtherHobbes12d ago

AI development has become an excuse for ignoring consent. Of course it's possible to filter out requests. But culturally with X, it's not remotely likely, unless compelled by regulation with teeth.

Retr0id12d ago

I can see how it'd be trivial to block known celebrities, but how do you handle everyone else?

rrr_oh_man12d ago

> trivial to block known celebrities

see here for one example: https://news.ycombinator.com/item?id=47370100

wolvoleo12d ago

Do you need to? It doesn't know everyone else. Or at least it shouldn't.

XorNot12d ago

I mean a realistic take is to simply not use source images containing people at all.

AIs have been able to invent fictional people longer then they've been able to modify existing images.

wolvoleo12d ago

You can just forbid using existing images as a source and describe them purely by text.

trollbridge12d ago

Portray fictional characters?

Retr0id12d ago

There are 8 billion humans, any fictional human is going to look almost exactly like at least one real human.

wolvoleo12d ago

Yes but for bullying purposes this is not useful. You're not going to try generating a pic 8 billion times till you get it right.

1 more reply

trollbridge12d ago

How about obviously fictional portrayals then? Somewhat cartoonish or anime or artistic etc

1 more reply

j / k navigate · click thread line to collapse

0 comments

_fizz_buzz_12d ago

nitwit00512d ago

Your filter has to pick out that, while they did not ask for a specific person, the practical result is likely to be the same. That's going to be tough to get near perfect.

TheOtherHobbes12d ago

AI development has become an excuse for ignoring consent. Of course it's possible to filter out requests. But culturally with X, it's not remotely likely, unless compelled by regulation with teeth.

Retr0id12d ago

I can see how it'd be trivial to block known celebrities, but how do you handle everyone else?

rrr_oh_man12d ago

> trivial to block known celebrities

see here for one example: https://news.ycombinator.com/item?id=47370100

wolvoleo12d ago

Do you need to? It doesn't know everyone else. Or at least it shouldn't.

XorNot12d ago

I mean a realistic take is to simply not use source images containing people at all.

AIs have been able to invent fictional people longer then they've been able to modify existing images.

wolvoleo12d ago

You can just forbid using existing images as a source and describe them purely by text.

trollbridge12d ago

Portray fictional characters?

Retr0id12d ago

There are 8 billion humans, any fictional human is going to look almost exactly like at least one real human.

wolvoleo12d ago

Yes but for bullying purposes this is not useful. You're not going to try generating a pic 8 billion times till you get it right.

1 more reply

trollbridge12d ago

How about obviously fictional portrayals then? Somewhat cartoonish or anime or artistic etc

1 more reply

j / k navigate · click thread line to collapse