Sir Paul McCartney backs calls to stop mass copyright theft by generative AI companies. www.theguardian.com/technology/2... - ThreadSky

ednewtonrex.bsky.social • 79 days ago

Sir Paul McCartney backs calls to stop mass copyright theft by generative AI companies.

https://www.theguardian.com/technology/2024/dec/10/paul-mccartney-ai-warning-uk-debates-copyright-laws

1 / 2

Comments

chemiwizard.bsky.social•79 days ago

I am willing to listen to artists speak about the danger of theft by AI. I am really not eager to hear from him in from of a Disney logo. Certain entities have abused the copyright laws and make any debate of the subject very difficult

ednewtonrex.bsky.social•79 days ago

Presumably it’s a file photo they have of him and not where he actually made the statement

chemiwizard.bsky.social•79 days ago

Oh I know, just not who i would want making this case. Metallica or those involved in fighting napster and suing college students over MP3s would be good faces for the debate either.

musicmark.bsky.social•79 days ago

Sure, he's rich

ednewtonrex.bsky.social•79 days ago

I don’t follow

tljamesa.com•79 days ago

Is there something points to what and how much has been scraped? Is it everything already? I don't think it lets ai take over, but it does reduce the value of everything a bit, right?

ednewtonrex.bsky.social•79 days ago

Suno, an AI music company, have admitted in court filings that the training data they’ve used “includes essentially all music files of reasonable quality that are accessible on the open Internet”.

tljamesa.com•79 days ago

What about text? I've heard people say "everything." Everything is a lot!

707kat.bsky.social•76 days ago

It is alot, but text do take up a whole lot less than an image or MP3. Still, there's also a whole lot more of it. If it's on the internet, assume it's been trained on.

And if it's not on the internet, it might have been put there by people for the purpose of training on it. Right now you see a-

707kat.bsky.social•76 days ago

whole lot of holding and investment companies selling the rights to train on their magazines and news to OpenAI amongst other things. Presenting it as a partnership. Assume these magazines have been trained on already.

Small publications and archives have been suffering something akin to-

707kat.bsky.social•76 days ago

ddos attacks from the likes of OpenAI, Anthropic, Bytedance etc. costing them money on top of ripping them off. OpenAI told a tiny website that they wouldn't pay them because they were too easy to rip off. So it's only the big companies who get deals like that. And not to the benefit of their staff.

Comments

Posting Rules

Reply