Elvis signs open letter against AI developers for using music without permission
-
sweetest punch
- Posts: 5986
- Joined: Sat Apr 03, 2004 5:49 am
- Location: Belgium
Elvis signs open letter against AI developers for using music without permission
Since you put me down, it seems i've been very gloomy. You may laugh but pretty girls look right through me.
Re: Elvis signs open letter against AI developers for using music without permission
Open letters are cool. I guess… but artists need to get their shit together and start creating their own tools to combat AI technology. Is there anyone in the music industry looking at ways to combat AI? I don’t know? Something tells me, no?
Not to get all radical here… but artists need to take technology in their own hands … maybe… Gee, I don’t know… Try a SQL injection command that might cripple the AI service that you feel is violating a copyright?
Technology has done a ton of damage to our socioeconomic condition and sadly our politicians are behind the curve in terms of how tech impacts our lives…
Here is a paper which appears to be written by academic types. It looks like prompt attacks, injections on AI generative models is the way to go if you want to save your copyrighted material - https://arxiv.org/pdf/2310.13828.pdf
To Note: The paper covers text-to-image models but there is probably a way to cover any AI generative model. The real issue is identifying the AI call at runtime.. when AI is invoked to call or create content... so runtime security can or may be able to block certain calls (content) to be used or if ingested..
To Note # 2: I'm a total clown/idiot. Above represents ideas from my feeble mind. Peace.
Not to get all radical here… but artists need to take technology in their own hands … maybe… Gee, I don’t know… Try a SQL injection command that might cripple the AI service that you feel is violating a copyright?
Technology has done a ton of damage to our socioeconomic condition and sadly our politicians are behind the curve in terms of how tech impacts our lives…
Here is a paper which appears to be written by academic types. It looks like prompt attacks, injections on AI generative models is the way to go if you want to save your copyrighted material - https://arxiv.org/pdf/2310.13828.pdf
To Note: The paper covers text-to-image models but there is probably a way to cover any AI generative model. The real issue is identifying the AI call at runtime.. when AI is invoked to call or create content... so runtime security can or may be able to block certain calls (content) to be used or if ingested..
To Note # 2: I'm a total clown/idiot. Above represents ideas from my feeble mind. Peace.