Current Events > Sarah Silverman sues OpenAI and Meta claiming AI training infringed copyright.

(+) Yeah! (+)

https://www.theguardian.com/technology/2023/jul/10/sarah-silverman-sues-openai-meta-copyright-infringement

Its more then just voice actors now...its going to get nasty as more and more demand something's done with AI.

I don't think courts are going to want to touch this.

They are going to have to at some point...its all if they have the guts to do it now or do it when things have gone fully to hell.

Necronmon posted...

They are going to have to at some point...its all if they have the guts to do it now or do it when things have gone fully to hell.

Courts could end up setting precedent which they might want to avoid for new things like this.

Fuck ai, and fuck its supporters and creators. AI should be illegal world wide, but we all know its really the next big thing for militarization.

Definitely needs to be heavily regulated. If identity theft is a crime, voice theft should be, too.

Is it moddable to discuss a deep fake?

Soon:
AI can only be trained with public domain assets, and ends up being absurdly racist.

"Open"Ai is the most Orwellian name for them

I think there's at least a good argument that the use of AI in these circumstances violates one's right of publicity, which is the right to exclusive economic use of a person's identity.

This is definitely a case where the law is far behind technology, though. Statutes will likely be needed to regulate since I don't think current IP law is able to address these issues.

Honeymoon phase with AI ended months ago and we're entering the complete utter contempt divorce phase now.

Google hath made them! The Courts shall break them!

Let us not stop at the various AI, though -- we shall next strike at the Power Looms, they just copy what weavers do! Basically infringing their skills in a copyright'ish manner! All Automation is the enemy!

Bring back the Switchboard operators!

Bring back the Bank Tellers!

Bring back the Pin Boys!

Bring back the Lift operators!

Bring back the Weavers!

Tyranthraxus posted...

I don't think courts are going to want to touch this.

They're gonna have to address the issue, it's not going away.

No need to go full Dune on it...just make sure that it can't make it so that the majority of humanity has no purpose other then being chew dolls for the" Gods" .

just because ai generated jokes suck doesnt mean its ripping off your act sarah

We need to make a safe space for AI. I am willing to move to another country if the laws in my country make it prohibitive to work with AI. Anyone who doesn't want it in their life are welcome to it.

shes a complete boomer karen

https://twitter.com/SarahKSilverman/status/1411908697501089803

UndefeatedGOAT posted...

shes a complete boomer karen

https://twitter.com/SarahKSilverman/status/1411908697501089803

Sarah is a racist piece of scum but what she says is right or you didn't understand what she said. Loud fireworks should be banned and events having fireworks should be only done privately with "fireworks" no louder or no more visible than those tiny popper thingys

I want AI to blow up, it's applications for research, business, and medical are fucking amazing in the future. Public use will inevitably be dumb shit but that's whatever. The courts need to settle the line between human and AI and do it quickly though, otherwise everything is gonna go to hell right quick.

BigB0ss13 posted...

Sarah is a racist piece of scum but what she says is right or you didn't understand what she said. Loud fireworks should be banned and events having fireworks should be only done privately with "fireworks" no louder or no more visible than those tiny popper thingys

thats a lot of shoulds, can you come to my area and Diwali and tell people to stop using fireworks and see how it goes

idk what it is with dogs though, mine doesnt care about firework but then is scared of other random shit

[LFAQs-redacted-quote]

If AI can't function properly within the confines of society and law that is a failing on the creators part. And if you think we'll have the same tools for long you're absolutely insane. Give it a year or two and they'll have premium paywalled AI models that outstrip publicly available ones hundreds of times over, at prices the public can't hope to pay on average.

This is going to be a repeat of when social media first got big. Silverman is basically claiming that her copyrighted material was used without her permission, but we have no idea where that data came from. If people are posting her specials, or transcripts of them, on public Facebook groups, and openai scrapes facebook, then who exactly is the responsible party for the copyright infringement? Facebook claims indemnity by arguing that they're not a publisher, they're just a forum. So it would be weird for openai to be responsible for material they scraped from a site that is claiming they're not responsible of copyrighted stuff ends up there.

SAG-AFTRA and WGA are both fighting AI within the film industry right now. Really, if SAG doesn't buckle, it's just a matter of time until execs give into their demands. AI is pretty much the top of the bill on their list now even if it isn't the most immediately important. If SAG goes to strike then it won't end until a deal restricting AI is made

Barber102 posted...

Fuck ai, and fuck its supporters and creators. AI should be illegal world wide, but we all know its really the next big thing for militarization.

this

Tyranthraxus posted...

Courts could end up setting precedent which they might want to avoid for new things like this.

They have years of experience with suits involving some one else's work.....i think they can handle it the same way with ai tools.

I'm totally against production companies and studios and shit using ai to replace people but I couldn't give less of a fuck about people uploading videos online of Metallica singing Taylor Swift or whatever the fuck

[LFAQs-redacted-quote]


We don't need AI for shit like faking voice acting using samples from actual work. How is that useful for the average person?

s0nicfan posted...

This is going to be a repeat of when social media first got big. Silverman is basically claiming that her copyrighted material was used without her permission, but we have no idea where that data came from. If people are posting her specials, or transcripts of them, on public Facebook groups, and openai scrapes facebook, then who exactly is the responsible party for the copyright infringement? Facebook claims indemnity by arguing that they're not a publisher, they're just a forum. So it would be weird for openai to be responsible for material they scraped from a site that is claiming they're not responsible of copyrighted stuff ends up there.

There's an obvious target here. Regulation on data scraping and where/when it can be used. It's been massively out of hand for a while and the only thing really hurt by it would be advertising agencies, and it wouldn't make an enormous dent in them.

Y'know, actually giving people a little fucking privacy on the Internet.

[LFAQs-redacted-quote]

Watch the issue only get addressed once a major production company or media group raises the issue. Who has more money to further their cause, media groups or artists? There's gonna be some bribery, I mean lobbying, going on.

Punished_Blinx posted...

We don't need AI for shit like faking voice acting using samples from actual work. How is that useful for the average person?

skyrim mods (and other games too i guess but who cares about those)
so many baller quest/companion mods with non existent or clearly amateur voice acting, or stitched samples
xvasynth sounds like shit unless you're really meticulous with it, but it was somewhat revolutionary in the modding scene

GeraldDarko posted...

Watch the issue only get addressed once a major production company or media group raises the issue. Who has more money to further their cause, media groups or artists? There's gonna be some bribery, I mean lobbying, going on.

There's also the existing laws around derivative works and fair use. If Sarah Silverman's special went into a model with 100 million other things, and that model produces something that she claims is hers, there's a real question to ask around whether what is generated is sufficiently changed from the original to count.

ArsGoetia posted...

skyrim mods (and other games too i guess but who cares about those)
so many baller quest/companion mods with non existent or clearly amateur voice acting, or stitched samples
xvasynth sounds like shit unless you're really meticulous with it, but it was somewhat revolutionary in the modding scene

Honestly the AI companions still sound like trash and are completely immersion braking unless you're completely brain rotted to only respond to memes, and if the quests are amazing they would be equally so without it... Otherwise they weren't amazing in the first place.

s0nicfan posted...

There's also the existing laws around derivative works and fair use. If Sarah Silverman's special went into a model with 100 million other things, and that model produces something that she claims is hers, there's a real question to ask around whether what is generated is sufficiently changed from the original to count.

That can already be a gray area, now throw AI into it. Who knows exactly how this will all play out.

R_Jackal posted...

AI companions still sound like trash and are completely immersion braking unless you're completely brain rotted to only respond to memes

Lighten up

R_Jackal posted...

Honestly the AI companions still sound like trash and are completely immersion braking unless you're completely brain rotted to only respond to memes, and if the quests are amazing they would be equally so without it... Otherwise they weren't amazing in the first place.

Babies are horrible at walking so we should just throw them away.

R_Jackal posted...

There's an obvious target here. Regulation on data scraping and where/when it can be used. It's been massively out of hand for a while and the only thing really hurt by it would be advertising agencies, and it wouldn't make an enormous dent in them.

Y'know, actually giving people a little fucking privacy on the Internet.

The issue isn't data scraping. It's fundamental ownership. Sticking with my facebook example, if the moment you post something online you consent to rescinding ownership, then if a company pays another company to scrape that data you technically have no right to complain, because that data was never yours in the first place. HOW they scrape the data is irrelevant.

GeraldDarko posted...

Lighten up

I think the fact that a lot of people have this mentality over serious issues these days is why the world is in the state it is in now.

Necronmon posted...

https://www.theguardian.com/technology/2023/jul/10/sarah-silverman-sues-openai-meta-copyright-infringement

Its more then just voice actors now...its going to get nasty as more and more demand something's done with AI.

i told you. This is inevitable.

s0nicfan posted...

The issue isn't data scraping. It's fundamental ownership. Sticking with my facebook example, if the moment you post something online you consent to rescinding ownership, then if a company pays another company to scrape that data you technically have no right to complain, because that data was never yours in the first place. HOW they scrape the data is irrelevant.

And that right there is why I never liked social media.

R_Jackal posted...

I think the fact that a lot of people have this mentality over serious issues these days is why the world is in the state it is in now.

AI and intellectual property are serious issues that aren't addressed by you suggesting that anyone that likes skyrims companions suffer from brain rot. C'mon... take it easy.

s0nicfan posted...

The issue isn't data scraping. It's fundamental ownership. Sticking with my facebook example, if the moment you post something online you consent to rescinding ownership, then if a company pays another company to scrape that data you technically have no right to complain, because that data was never yours in the first place. HOW they scrape the data is irrelevant.

wdym the issue

isnt scraping bad for the environment as well? From a sarah silverman perspective people could be reposting her shit illegally and it got scraped

R_Jackal posted...

And that right there is why I never liked social media.

*shrugs* terms of service. There's a reason people say "if they don't charge you, you are the product." The entire backbone of the internet as a community is based on taking advantage of massive private industry tools developed at no cost to the consumer. Would you use facebook if your data was private but it was $100/yr? For most people, probably not.

Kill data ownership as a concept for companies, and social media, all of it, vanishes overnight.

But this isn't some big secret. Go back to 2018 and people were openly declaring that "data is the new oil": https://www.auvsi.org/events/webinars/auvsi-webinar-data-new-oil-and-other-xponential-ideas

Idk, fair being fair, a voice actor, or the like, should have more of a say than a random nobody/leach.

In regards to what happens to their likeness. So to speak.

Not asinine to think such people deserve such freedoms. Especially if it's their livelihood.

It's not like I'm against ai voices. But create your own shit. Not a complicated concept imo.

UndefeatedGOAT posted...

wdym the issue

isnt scraping bad for the environment as well? From a sarah silverman perspective people could be reposting her shit illegally and it got scraped

I mean what the legal issue is. Banning scraping doesn't mean anything since companies can package and sell that data (they they legally own) through other means. The only person who would be seriously hurt by data scraping legislation are university researchers that can't afford the API pipeline costs these companies charge for alternative access.

And IIRC scraping isn't bad for the environment. You're probably thinking of mining which is a crypto thing.

s0nicfan posted...

I mean what the legal issue is. Banning scraping doesn't mean anything since companies can package and sell that data (they they legally own) through other means. The only person who would be seriously hurt by data scraping legislation are university researchers that can't afford the API pipeline costs these companies charge for alternative access.

And IIRC scraping isn't bad for the environment. You're probably thinking of mining which is a crypto thing.

i suppose its not too much worse than me googling youtube, and google searching millions of pages, sure

I think the complaint can be valid if illegally posted content is scraped and forever part of the AIs database

s0nicfan posted...

*shrugs* terms of service. There's a reason people say "if they don't charge you, you are the product." The entire backbone of the internet as a community is based on taking advantage of massive private industry tools developed at no cost to the consumer. Would you use facebook if your data was private but it was $100/yr? For most people, probably not.

Kill data ownership as a concept for companies, and social media, all of it, vanishes overnight.

But this isn't some big secret. Go back to 2018 and people were openly declaring that "data is the new oil": https://www.auvsi.org/events/webinars/auvsi-webinar-data-new-oil-and-other-xponential-ideas

I don't use it anyway, and it's always been crazy to me that people give it away like candy. If someone wants my information, they should pay me.

UndefeatedGOAT posted...

i suppose its not too much worse than me googling youtube, and google searching millions of pages, sure

I think the complaint can be valid if illegally posted content is scraped and forever part of the AIs database

Sure, but that leads back to my first post around who is responsible. Is it the person posting illegal content to public forums, the forum that took no action to remove it, the company that developed the scraping tools for not having a means of checking, or openai itself for using that data without checking every single item for legality?

EDIT: Which is all to say this is an extremely complicated, highly interconnected series of laws, terms, ownership, and derivative works all bundled together.