Comment by UmYeahNo | Hacker News Reader

Comment by UmYeahNo | Hacker News Reader

Parent Full thread

>But remember, the bias ends with the user using the tool. If I want a black scientist, I can just say "black scientist".

That is a really, really, narrow viewpoint. I think what people would prefer is that if you query "Scientist" that the images returned are as likely to be any combination of gender and race. It's not that a group is "fragile", it's that they have to specify race and gender at all, when that specificity is not part of the intention. It seems that they recognize that querying "Scientist" will predominantly skew a certain way, and they're trying in some way to unskew.

Or, perhaps, you'd rather that the query be really, really specific? like: "an adult human of any gender and any race and skin color dressed in a laboratory coat...", but I would much rather just say "a scientist" and have the system recognize that anyone can be a scientist.

And then if I need to be specific, then I would be happy to say "a black-haired scientist"

Have you seen the queries that are used to generate actually useful results rather than just toy demonstrations? They look a lot more like your first example except with more specificity. It'd be more like "an adult human of any gender and any race and skin color dressed in a laboratory coat standing by a window holding a beaker in the afternoon sun. 1950s, color image, Canon 82mm f/3.6, desaturated and moody." so if instead you are looking for an image with a person of a specific ethnicity or gender then you are for sure going to add that in along with all of the details. If you are instead worried about the bias of the person choosing the image to use then there is nothing short of restricting them to a single choice that will fix that and even in that case they would probably just not use the tool since it wasn't satisfying their own preferences.

This is a problem with generative models across the board. It's important that we don't skew our perceptions by GAN outputs as a society, so it's definitely good that we're thinking about it. I just wish that we had a solution that solved across the class of problems "Generative AI feeds into itself and society (which is in a way, a generative AI), creating a positive feedback loop that eventually leads to a cultural freeze"

It's way bigger than just this narrow race issue the current zeitgeist is concerned about.

But I agree, maybe I should skew to being optimistic that at least we're trying

Kind of funny that NN tech is supposed to construct some upper dimensional understanding, yet realistically cannot be expected to be able to generate gender and race indeterminate portrayal of a scientist.

Reply on news.ycombinator.com