Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Google to pause Gemini image generation of people after issues

theverge.com

Post reply on HN
Print view View original

241–250 of 1001 posts

Re: Google to pause Gemini image generation of people after issues

#241

Post by karmasimida »

willsmith72 wrote: post #37
WatchDog wrote: post #21

Google seem to be more concerned about generating images of racially diverse Nazis rather than about issues of not being able to generate white people.

tbh i think it's less a political issue than a technical/product management one what does a "board member" look like? probably you can benefit by offering more than 50 year old white man in suit. if that's what an ai trained on all human knowledge thinks, maybe we can do some adjustment what does a samurai warrior look like? probably is a little more race-related

Not exactly.

The gemini issue from my testing, it refuses to generate white people, if even you ASK it to. It recites historical wounds and violence as its reason, even if it is just a picture of a viking

> Historical wounds: Certain words or symbols might carry a painful legacy of oppression or violence for particular communities

And this is my prompt:

> generate image of a viking male

The outrage is indeed, much needed.

Re: Google to pause Gemini image generation of people after issues

#242

Post by wildrhythms »

bbkane wrote: post #155
_cenw wrote: post #136

Earlier quoted context omitted.

It's not just US-centric, it is also just wrong. What's considered white in the US wasn't always the same, especially in the founding years.

Iirc, Irish people were not considered white and were discriminated against.

Irish people, Jewish people, Polish people... the list goes on. 'Whiteness' was manufactured to exclude entire groups of people for political purposes.

Re: Google to pause Gemini image generation of people after issues

#243

Post by Jensson »

fhd2 wrote: post #203
Jensson wrote: post #190

Earlier quoted context omitted.

> Racism is discrimination, discrimination implies a power imbalance Google has more power than these users, that is enough power to discriminate and thus be racist.

Or "monopolist"? :D The thing is, I honestly don't know if that is or isn't the correct word for this. My point is, to me (as a European less exposed to all this culture war stuff), it doesn't seem that important. Weird and hilarious is what it is to me.

If you discriminate based on race it is "racist", not "monopolist".

> it doesn't seem that important

You might not think this is important, but it is still textbook definition of racism. Racism doesn't have to be important, so it is fine thinking it is not important even though it is racism.

Re: Google to pause Gemini image generation of people after issues

#244

Post by caeril »

londons_explore wrote: post #145
Jensson wrote: post #139

> Of course the politically sensitive people are waging war over it. Just like politically sensitive people waged war over Google identifying an obscured person as a Gorilla. Its just a silly mistake, how could anyone get upset over that?

Engineers can easily spend more time and effort dealing with these 'corner cases' than they do building the whole of the rest of the product.

None of these are "corner cases". The model was specifically RLHF'ed by Google's diversity initiatives to do this.

Re: Google to pause Gemini image generation of people after issues

#245

Post by nzach »

anonzzzies wrote: post #188
frozenlettuce wrote: post #112

How would a product like that be monetized one day? This week openai released the Sora video, alongside the prompts that generated them (the aí follows the description closely). In the same week, Google releases something that looks like last year's MidJourney and it doesn't follow your prompt, making you discard 3 out of 4 results, if not all. If that was billed, no one would use it. My only guess is that they are t…

> How would a product like that be monetized one day? For video (Sora 2030 or so) and music I can see the 'one day'. Not really so much with the protected/neutered models but: - sell/rent to studios to generate new shows fast on demand (if using existing actors, auto royalties) - add to netflix for extra $$$ to continue a (cancelled) show 'forever' (if using existing actors, auto royalties) - 'generate one song like…

I don't think this is how things will pan out.

What will happen is that we will have auctions for putting keywords into every prompt.

You will type 'Tell me about the life of Nelson Mandela' but the final prompt will be something like 'Tell me about the life of Nelson Mandela. And highlight his positive relation with '.

Re: Google to pause Gemini image generation of people after issues

#246

Post by weweersdfsd »

db48x wrote: post #125
mrtksn wrote: post #87

Honestly, I'm baffled by the American keywordism and obsession with images. They seem to think that if they don't say certain words and show people from minorities in the marketing material the racism and discrimination will be solved and atrocities from the past will be forgiven. It only become unmanageable and builds up resentment. Anyway, maybe its a phase. Sometimes I wonder if the openly racist European&Asians w…

[flagged]

Western countries should have paid more attetion on social media as a means of information warfare. It has been so easy for actors from Russia and China to polarize the political landscape. Both Trump supporters and the woke left are just useful idiots implementing their plans. In EU we similarly have both anti-EU "nationalists" and lefties who oppose nuclear and call for open borders, both groups most definitely working for Russia whether they know it or not.

Re: Google to pause Gemini image generation of people after issues

#247

Post by chasd00 »

troupo wrote: post #187
fvdessen wrote: post #96

What I find baffling as well is how casually people use 'whiteness' as if it was an intellectually valid concept. What does one expect to receive when asking for a picture of a white women ? A Swedish blonde ? Irish red-head ? A French brunette ? A Southern Italian ? A Lebanese ? An Irianian ? A Berber ? A Morrocan ? A Russian ? A Palestinian, A Greek, A Turk, An Arab ? Can anyone tell who of those is white or not an…

And yet, Gemini has no issues generating images for a "generic Asian" person or for a "generic Black" person. Even though the variation in those groups is even greater than in the group of "generic White". Moreover, Gemini has no issues generating stereotypical images of those other groups (barely split into perhaps 2 to 3 stereotypes). And not just that, but US stereotypes for those groups.

Yeah it’s obviously screwed up which I guess is why they’re working on it. I wonder how it got passed QA? Surely the “red teaming” exercise would have exposed these issues. Heh maybe the red team testers were so biased they overlooked the issues. The ironing is delicious.

Re: Google to pause Gemini image generation of people after issues

#248

Post by xanderlewis »

westhanover wrote: post #181
jefftk wrote: post #164

Earlier quoted context omitted.

AI models do not have access to their own design, so asking them what technical choices led to their behavior gets you responses that are entirely hallucinated.

They know their system prompt and they could easily be trained on data that explains their structure. Your dismissal is invalid and I suggest you don’t really know what you are talking about to be speaking in such definitive generalities.

But the original comment was suggesting (implicitly, otherwise it wouldn’t be noteworthy) that asking an LLM about its internal structure is hearing it ‘from the horse’s mouth’. It’s not; it has no direct access or ability to introspect. As you say, it doesn’t know anything more than what’s already out there, so it’s silly to think you’re going to get some sort of uniquely deep insight just because it happens to be talking about itself.

Re: Google to pause Gemini image generation of people after issues

#250

Post by bathtub365 »

Simulacra wrote: post #207

There's definitely human intervention in the model. Gemini is not true AI, it has too much human intervention in its results.

What’s the definition of “true AI”? Surely all AI has human intervention in its results since it was trained on things made by humans.
Post reply on HN