Hacker News Reader

Comment by xepriot | Hacker News Reader

What in the world is going on with these convoluted framings? Just call it racism - not a 'shortcut' or an 'overcorrection'. Why not just clearly, simply, and precisely describe what is going on? This is anti-white racism. Google Gemini was erasing white people from history and also the present, and when it wasn't simply erasing them, it was diluting them - going partway towards erasure.

What I've said is bare, basic truth, and not ideological. If it offends or provokes factual or semantic quibbling, then such reaction indicates pretzel thinking about moral good and 'racism' in the minds of people like Smith.

View on HN

The convoluted framings come from people who want to actually persuade those who don't already agree with them. If you want someone to listen to you you have to first meet them where they are, not because you're worried about offense but because you want them to change their mind more than you want to state "bare, basic truth".

The people behind Gemini's "diversity" sincerely believe they're doing the right thing. Calling them racist doesn't help change their mind, it just gets you filed under "reactionary" and ignored. This piece takes a more effective approach by saying "I'm with you, I think you want to do the right thing, but here's why what you're doing is actually counterproductive".

It's an effective rhetorical strategy for the target audience, you're just not that audience because you already see the problem.

This is certainly charitable to Noah Smith, but I think it's more likely he earnestly has the overcomplicated understanding of racism he appears to demonstrate in his article. And it does have the same effect as if he did: he advances the confused understanding of racism as something that only happens to non-whites.

Of course everyone has good intentions and likes to be flattered. It's possible to flatter with tone, and without dissembling.

Yes, in order to convincingly say "I'm with you, I think you want to do the right thing" you do have to actually believe that America has a race problem that needs to be addressed and you have to believe that the people taking shortcuts sincerely believe they're doing the right thing. So yeah, Noah Smith does presumably hold that understanding of racism.

I wasn't trying to say that he was lying for the sake of the article, I was trying to say that he's framing it in a way that is intended to be persuasive to a particular group of people that you're not a part of. If you pressed him on it in a private conversation I wouldn't be surprised if he acknowledged that the behaviors involved here are racist (by some definition) even though the people are trying to be the opposite. But he knows better than to level that word at his audience.

I disagree. You just have to be committed to doing the right thing. If some ideological complex or other is opposed to this, you can compassionately tell people the truth, writing under the assumption - which you may emphasise as much as you like - that they are Good People who also want to do the right thing.

I would tell him that however he justifies concealing or dressing up his beliefs, such dissembling really does have the effect of increasing confusion and rancor - he is, of course, also a Good Person, but this is not a good tactic.

His problem isn't that he overcomplicated a theory, his problem is personal and he expresses it with racism.

> diverse [...] which the AI interpreted as meaning “nonwhite”

While an AI isn't objective in any way, it is the result of the dataset. This is also an accusation against the DEI groups that grant themselves the title of being less biased or just self-reflective enough. Problem is that they simply are not, on the contrary.

Google overengineers itself over and over again, liberal managers are dangerous. Make generative AI that is grounded in historical facts and let people play with it. For example if all popes were white, generate white pope image but let me also generate black pope because why not, if technology can do it.

Because there's a newspeak movement to try to dilute the definition of 'racism' as something that can't be perpetrated against relatively powerful racial groups.

Clearly, you've never spent any time on Metafilter. There's been a long push to reformulate the concept of racism such that white people cannot be a victim of it. Even in an environment with a reversal of the usual population statistics. Similarly, cannot be sexist against men, and so on.

Right now, the DEI exists in these models at a level where context is not considered. Not for time or locale, nor for, say, the correlation between sex and different populations present in a given profession. It's just an unconsidered DEI goo sprayed over the outputs.

Is it 'racism' when facial recognition algorithms perform more poorly on dark skin?

It seems like a technical glitch to me. Similarly here, they want the models to represent ethnically diverse individuals, but the model doesn't understand historical context well. It's a technical shortcoming.

> Is it 'racism' when facial recognition algorithms perform more poorly on dark skin?

No, because that occurred due to an oversight of not including a representative set of training data. At worst it's implicit racism due to oversight and ignorance.

What Google seemed to do here was more explicit racism, by setting configuration in their system that directly and intentionally discriminates based on race. It's fundamentally worse.

They both seem like goofs to me. I don't see much qualitative distinction. In both cases, engineers have produced racially dubious results by getting sloppy or lazy with their AI models. Personally, I wouldn't call that "racism" but I see how one could make that argument.

If you read the OP you will find a link to Google statement where Google officially admits they did this intentionally. They tried to shoot an arrow, and they "missed the mark", oops. They're not sorry for shooting the arrow, they are sorry for missing the mark.

Nobody who worked on facial recognition that failed to recognize black people, nobody in that team was intentionally trying to fail on black people. They didn't try to shoot an arrow, although they accidentally shot one in their own leg.

See how these things are different?

Is it not obvious that the distinction is intent to discriminate based on race?

Google, in a self imposed noble desire to rectify poor representation of training data, intentionally injected race discrimination in their system. This intent is what makes it fundamentally worse than unintentionally having non diverse training data.

They created a mirror. They didn’t like that the mirror showed reality, so they made it into a one way race-changing funhouse mirror.

Neither case is a purely technical shortcoming—in both the case of facial recognition algorithms performing poorly on dark skin and the case of Gemini refusing to produce images of white people, the technology works that way because of the priorities of the people developing the software, and we can derive the developers' priorities from the software's limitations.

A piece of facial recognition software that doesn't do well with dark skin shows that the team that built it either didn't think to test it on dark skin or didn't feel it was important to get dark skin working before release.

An image generation model that both doesn't produce images of white people spontaneously and will actively refuse to do so when asked shows that the team that built it either failed to ever ask it to produce images of white people or believed that it was a good thing for their model to refuse to produce such images.

In neither case is the problem purely technical.

It could be a consequence of racism. Racism lead to white people dominating the United States, which lead to images of white people dominating the internet, which lead to images of white people dominating the training data for facial recognition algorithms.

But it could also just reflect the composition of the population of the countries that dominate the internet.

No, when unintentioned outcomes occur in the world, this is not racism. The sun is not racist for burning white skin more severely than dark skin, and when AI are trained on disproportionately white data, their outputs are not racism.

When racist employees of Google intentionally racially cleanse their products, this is racism.

Let me know what is confusing about the concepts of agency and intention., and I will try to clarify

You appear to only comment here on HN when race comes up.

So, are the Caucasian people who were involved in any of these miss-optimizations, aka mistakes, racist against themselves? Are they traitors?

At the least they are oikophobes.

"I want a model that recognizes faces, but I didn't train it on enough black faces" => not-racism, according to you.

"I want a model that produces racially diverse images of people, but I didn't test it in racially homogenous historical contexts" => yes-racism, according to you.

> "I want a model that produces racially diverse images of people, but I didn't test it in racially homogenous historical contexts"

You are again misrepresenting the OP. If you go actually read the OP you will find it very clearly describes Google's initial steps towards fixing the problem, which was: double down on race-swapping in historical contexts. The OP very clearly (and correctly) argues that this is not the action of a person who goes "oops forgot to test in this context" but is rather the action of a person who very intentionally wanted the image generator to behave this way in this context.

Did the developers intend their product to manipulate race in its outputs or not? It's a simpler question than these contorted reformulations you present.

But probably not simple enough to pierce ideological bias. I'll ask instead: you really believe that it wasn't tested in historical and other contexts where white people were erased?

> I'll ask instead: you really believe that it wasn't tested in historical and other contexts where white people were erased?

Occam's Razor: Of course it wasn't or they wouldn't have released such a ridiculous embarrassing product! C'mon, I think you're the one wearing blinders here.

We'll see what comes out later, but I give it 90:10 odds that this behavior was seen in testing and accepted by developers/DEI commissars - the failure was in not forseeing the reaction of the non-DEI-extremist public.

Of a piece with 'Google more or less explicitly said I won't be promoted because I'm White' stories.

> Of course it wasn't or they wouldn't have released such a ridiculous embarrassing product!

Then please do explain why their first reaction was to double down on "diversity" in historical contexts, before eventually retreating?

Because adding the word "diverse" to the text of to their output is a lot easier than retraining their model? They're always going to try for the easy fix first.

Have you never tried tweaking the error text first before giving up and redesigning your form that's confusing users?

Hmmh. Fair enough. I didn't think about it from the perspective of "error text", but now that I think in that mental model, I think you're right here.

No, this is not limited to gemini.

Reply on news.ycombinator.com