> alleges that Carlin’s copyrighted materials and likeness were used without permission or appropriate licenses
It sounds like Carlin's estate doesn't understand how AI works. If it's illegal to use Carlin's material in this way, then all the popular AI models must be illegal as well lol
This exact question is currently being litigated, and using copyrighted material when training an LLM or AI model without permission could very well be illegal.
- The qualitative differences. You can't run an unbounded # of simultaneous and 24/7 instances of a natural human intelligence.
- Machine learning models are not people and do not have rights.
- Obtaining the training material itself requires massive copyright infringement. GPT-3 was even openly trained on books from a literal pirate torrent tracker.
> OpenAI hasn’t disclosed the datasets that ChatGPT is trained on, but in an older paper two databases are referenced; “Books1” and “Books2”. The first one contains roughly 63,000 titles and the latter around 294,000 titles
More likely the question is going to be how far parody and satire can allow for the use celebrity likeness, a much blurrier line and not anywhere near cut and dried; you only have to look at sketch comedy to see a whole lot of impressions where name, likeness and voice were all imitated for the purposes of comedy. Will Sasso himself did a bunch of that while he performed on MadTV.
It also has little to with AI. The fidelity of the imitation is higher, but is that the biggest issue when the content in question makes it plainly obvious that it's an impression? No one was fooled that it was Carlin, they were fooled that it was written by an AI.
If the content of the music is filled with commentary and jokes about the use of AI to create facsimiles of celebrities and comedians and the exploitation of the technology for the purposes of mass commercialization by billionaires, you might be able to.
After all, you can open a store selling shitty coffee and market it as Dumb Starbucks.
9 comments
[ 4.1 ms ] story [ 52.0 ms ] threadIt sounds like Carlin's estate doesn't understand how AI works. If it's illegal to use Carlin's material in this way, then all the popular AI models must be illegal as well lol
- The qualitative differences. You can't run an unbounded # of simultaneous and 24/7 instances of a natural human intelligence.
- Machine learning models are not people and do not have rights.
- Obtaining the training material itself requires massive copyright infringement. GPT-3 was even openly trained on books from a literal pirate torrent tracker.
What's stopping a human from reading all of the same books?
Even the most dedicated readers will only read a few thousand books in their lifetime. https://www.mentalfloss.com/article/93594/heres-how-many-boo...
ChatGPT is trained on orders of magnitude more books than a human can consume.
https://torrentfreak.com/authors-accuse-openai-of-using-pira...
> OpenAI hasn’t disclosed the datasets that ChatGPT is trained on, but in an older paper two databases are referenced; “Books1” and “Books2”. The first one contains roughly 63,000 titles and the latter around 294,000 titles
More likely the question is going to be how far parody and satire can allow for the use celebrity likeness, a much blurrier line and not anywhere near cut and dried; you only have to look at sketch comedy to see a whole lot of impressions where name, likeness and voice were all imitated for the purposes of comedy. Will Sasso himself did a bunch of that while he performed on MadTV.
It also has little to with AI. The fidelity of the imitation is higher, but is that the biggest issue when the content in question makes it plainly obvious that it's an impression? No one was fooled that it was Carlin, they were fooled that it was written by an AI.
I can't put out an album of my shitty music and market it as a new Michael Jackson album.
Putting "AI" in front of "Michael" doesn't change anything.
After all, you can open a store selling shitty coffee and market it as Dumb Starbucks.