OpenAI now tries to hide that ChatGPT was trained on copyrighted books, including J.K. Rowling’s Harry Potter series::A new research paper laid out ways in which AI developers should try and avoid showing LLMs have been trained on copyrighted material.
An LLM is not a person, it is a product. It doesn’t matter that it “learns” like a human - at the end of the day, it is a product created by a corporation that used other people’s work, with the capacity to disrupt the market that those folks’ work competes in.
And it should be able to freely use anything that’s available to it. These massive corporations and entities have exploited all the free spaces to advertise and sell us their own products and are now sour.
If they had their way they are going to lock up much more of the net behind paywalls. Everybody should be with the LLMs on this.
You are somehow conflating “massive corporation” with “independent creator,” while also not recognizing that successful LLM implementations are and will be run by massive corporations, and eventually plagued with ads and paywalls.
People that make things should be allowed payment for their time and the value they provide their customer.
People are paid. But they’re greedy and expect far more compensation then they deserve. In this case they should not be compensated for having an LLM ingest their work work if that work was legally owned or obtained
How are we going to make ai, if it can’t learn?
First, we don’t have to make AI.
Second, it’s not about it being unable to learn, it’s about the fact that they aren’t paying the people who are teaching it.
yeah lets not explore this technology because it might hurt some copyrights holders
LOOOOL fuck em
FTFY. Corporations shouldn’t be making a fucking dime from any of these works without fairly paying the creators.
Then give the AI a library card, feel better?
The reasoning that claims training a generative model is infringing IP would still mean a robot going into a library with a card it has to optically read all the books there to create the same generative model would still be infringing IP.
Same way that counting cards is illegal
Counting cards isn’t illegal though lol
Exactly and neither is this.
Exactly
Humans can judge information make decisions on it and adapt it. AI mostly just looks at what is statistically what is most likely based on training data. If 1 piece of data exists, it will copy, not paraphrase. Example was from I think copilot where it just printed out the code and comments from an old game verbatim. I think Quake2. It isn’t intelligence, it is statistical copying.