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

TikTok is being flooded with racist AI videos generated by Google's Veo 3

arstechnica.com

Post reply on HN
Print view View original

51–60 of 98 posts

Re: TikTok is being flooded with racist AI videos generated by Google's Veo 3

#51

Post by jrflowers »

drdaeman wrote: post #22
CharlesW wrote: post #18

Earlier quoted context omitted.

> It isn’t even a “real” racism… Generating and distributing racist materials is racist regardless of the intent, even if the person "doesn't mean it". Simple thought experiment: If the content was CSAM, would you still excuse the perpetrators as victims of perversely incentivized attention economics?

I agree, but I believe the intent matters if we’re trying to identify why this happens. Racism is just less legally dangerous. There would be people posting snuff or CSAM videos, would that “sell”. Make social networks tough on racism and it’ll be sexism next day. Or extremist politics. Or animal abuse. Or, really, anything, as long as people strongly react to it. But, yeah, to avoid any misunderstanding - I didn’t m…

>it’s not the real issue here

I like this reasoning. “Trolling” is when people post things to irritate or offend people, so if you see something that’s both racist and offensive then it’s not really racist. If you see somebody posting intentionally offensive racist stuff, and you have no other information about them, you should assume that the offensiveness of their post is an indicator of how not racist they are.

Really if you think about it, it’s like a graph where as offensiveness goes up the racism goes down becau

Re: TikTok is being flooded with racist AI videos generated by Google's Veo 3

#53

Post by jrflowers »

The interesting thing about this is that it is the use case for these video generators. If the point of these tools is to churn out stuff to drive engagement, and the best way to do that is through content that is inflammatory, offensive, or misinformation, then that’s the ideal use case for them. That’s what the tool is for.

Re: TikTok is being flooded with racist AI videos generated by Google's Veo 3

#54

Post by accoil »

RockRobotRock wrote: post #41
defrost wrote: post #35

Earlier quoted context omitted.

Off the cuff the closest example to mind is the Paedogeddon!! special episode of the Brass Eye series created by Chris Morris. Admittedly that didn't satirize CSAM material, rather it cut hard into the reflexive reaction people have at the very thought of CSAM and peodophiles. https://en.wikipedia.org/wiki/Paedogeddon Moreover, that took a human to thread that needle, it'll be a while before AI generation can pass th…

This is the one thing we didn't want to happen

AI creating satire about media spreading hysteria? I don't think it's at that point.

Re: TikTok is being flooded with racist AI videos generated by Google's Veo 3

#55

Post by IAmGraydon »

SchemaLoad wrote: post #23
ivape wrote: post #17

I think it's fine to fingerprint AI generated images/videos. It's a massive privacy violation but I just can't see any other way. Too many people have always been and will always be unethical.

I've been wondering if ChatGPT makes such excessive use of EM dash just so people can easily identify AI generated content. Google wouldn't even need a fingerprint, they could just look up from their logs who generated the video.

The em-dash is one marker, but I’ve read that most LLMs create small but statistically detectable biases in their output to help them avoid reingesting their own content.

Re: TikTok is being flooded with racist AI videos generated by Google's Veo 3

#56

Post by ghushn3 »

prvc wrote: post #36

None of the examples shown in the video are passable hoaxes. They are all obvious burlesque-style parodies, albeit made in bad taste. They all also have clear and prominent hallmarks of AI generation. Anyone fooled by these has got bigger, prior problems than any potential belief instilled by these videos.

The problem is not that they are fooling anyone. No one thinks a woman is marrying a chimpanzee. The problem is that the videos are obviously and openly racist and being spread quite brazenly.

If I have to encounter a constant barrage of shitty racist (or sexist, or homophobic, or whatever) material just to exist online, I'm going to pretty quickly feel like garbage. (If not feel unsafe.) Especially if I'm someone who has other stressors in their life. Someone who is doing well, their life otherwise together, might encounter these and go, "Fucking idiots made a racist video, block."

But empathize with someone who is struggling? Who just worked 18 hours to make ends meet to come home and feed their kids and pay rent for a shitty apartment that doesn't fit everyone, and their kid comes up to them asking what this video means, and it just... gets past all their barriers. It wedges open so many doubts.

This isn't harmless.

Re: TikTok is being flooded with racist AI videos generated by Google's Veo 3

#57

Post by ghushn3 »

linotype wrote: post #52

People have to stop watching this trash. And I mean all of TikTok.

Some of TikTok is great. I mean, most of it is just dopamine hits, and it's potentially quite bad from a health perspective. But also, plenty of TikTok is news, or political theory, or thoughtful commentary, or explanations of how things work.

It's a bowl of fun size candy bars, with a few razors, a few drugs, a few rotten apples, etc. mixed in. You can, by and large, get the algorithm to serve you nothing but the candy, but you are still eating only candy bars at that point.

Some people can say no to infinite candy. Other people, like myself, cannot and it's a real problem.

Re: TikTok is being flooded with racist AI videos generated by Google's Veo 3

#58

Post by ghushn3 »

unsnap_biceps wrote: post #42
nvch wrote: post #20

The question is, who is acting in a racist manner here: the LLM that does what it can, or the humans sharing those videos?

Until we get a LLM that actually "thinks", it's just a tool like photoshop. Photoshop isn't racist if someone uses it to create racist material, so a LLM wouldn't be racist either.

I saw (on HN, actually) an academic definition for prejudice, discrimination, and racism that stuck with me. I might be butchering this a bit, but prejudice is basically thinking another group is less than purely because of their race. Discrimination is acting on that belief. Racism is discrimination based on race, particularly when the person discriminated against is a minority/less powerful person.

LLMs don't think, and also have no race. So I have a hard time saying they can racist, per se. But they can absolutely produce racist and discriminatory material. Especially if their training corpus contains racist and discriminatory material (which it absolutely does.)

I do think it's important to distinguish between photoshop, which is largely built from feature implementation ("The paint bucket behaves like this", etc.), and LLMs which are predictive engines that try to predict the right set of words to say based on their understanding of human media. The input is not some thoughtful set of PMs and engineers, it's "read all this, figure out the patterns". If "all this" contains racist material, the LLM will sometimes repeat it.

Re: TikTok is being flooded with racist AI videos generated by Google's Veo 3

#59

Post by hombre_fatal »

eurleif wrote: post #32

Somewhat related, on YouTube, there's a channel filled with fake police bodycam videos. The most-viewed of these are racially inflammatory, e.g.: https://www.youtube.com/watch?v=5AkXOkXNd8w The description of the channel on YouTube claims: "In our channel, we bring you real, unfiltered bodycam footage, offering insight into real-world situations." But then if you go to their site, https://bodycamdeclassified.com/ , w…

Dang, police bodycam videos are my guilty pleasure when I'm working out and just want dumb stimulation to pass the grind.

Definitely have watched enough videos from this channel to recognize its name. :(

Re: TikTok is being flooded with racist AI videos generated by Google's Veo 3

#60

Post by Dig1t »

CharlesW wrote: post #18
drdaeman wrote: post #9

> nothing drives engagement on social media like anger and drama There. It isn’t even a “real” racism, it’s more of a flamebait, where the more outrageous and deranged a take is, the more likely it would captivate attention and possibly even provoke a reaction. Most likely they primarily wanted to earn some buck from viewer engagement, and didn’t care about the ethics of it. Maybe they also had the racist agendas, ma…

> It isn’t even a “real” racism… Generating and distributing racist materials is racist regardless of the intent, even if the person "doesn't mean it". Simple thought experiment: If the content was CSAM, would you still excuse the perpetrators as victims of perversely incentivized attention economics?

The creation of CSAM is a crime because an underage person must be harmed in its creation by definition. Making an AI video of an offensive stereotype does not harm anyone in its creation. It is textbook free speech.

Clutch your pearls as much as you want about the videos, but forcibly censoring them is going to cause you to continue to lose elections.

Post reply on HN