“Established copyright doctrine will dictate that the Times cannot prevent AI models from acquiring knowledge about facts, any more than another news organization can prevent the Times itself from re-reporting stories it had no role in investigating,” OpenAI writes.
Oh boy, their defense is that their advanced predictive text can acquire knowledge? Please, proceed.
It still is defensible. I can quote a whole bunch of lines from “talladega nights” and “old school” verbatim. I can sing the entirety of “Amish paradise”, with close to 100% accuracy.
My recall ability does not mean that I’ve violated copyright.
This doesn’t matter. You personally reciting movie quotes as a private individual is fair use. OpenAI’s ChatGPT has a commercial purpose, and you could say it does compete with The NY Times.
Only indirectly - as in airplanes competing with cars. And the law generally encourages that type of competition as it leads to substantial innovation and economic growth.
That’s like saying amazon and mom and pop gift shops don’t compete. Like yeah, a lot of people will still prefer the atmosphere and curation of the mom and pop shop but that doesn’t fucking matter when the vast majority of people just use Amazon, driving the shop out of business. This despite the fact that Amazon is more general and only competes indirectly.
I could totally start a website, maybe call it “New York Stories”, read every single NYT article and then (working off my own memory, not copy/pasting the text) write the same story again and publish it. That would not be copyright infringement. In fact the NYT themselves do it all the time, publishing things that were originally reported elsewhere.
Do you have paying customers that ask you for movie scripts and song lyrics like OpenAi does? If so, the above would be flat out copyright infringement.
Oh boy, their defense is that their advanced predictive text can acquire knowledge? Please, proceed.
It would be a plausible defense if the AI model wasn’t regurgitating Times articles verbatim.
It still is defensible. I can quote a whole bunch of lines from “talladega nights” and “old school” verbatim. I can sing the entirety of “Amish paradise”, with close to 100% accuracy.
My recall ability does not mean that I’ve violated copyright.
This doesn’t matter. You personally reciting movie quotes as a private individual is fair use. OpenAI’s ChatGPT has a commercial purpose, and you could say it does compete with The NY Times.
Only indirectly - as in airplanes competing with cars. And the law generally encourages that type of competition as it leads to substantial innovation and economic growth.
That’s like saying amazon and mom and pop gift shops don’t compete. Like yeah, a lot of people will still prefer the atmosphere and curation of the mom and pop shop but that doesn’t fucking matter when the vast majority of people just use Amazon, driving the shop out of business. This despite the fact that Amazon is more general and only competes indirectly.
Are you charging for your performances?
No, but you’re not trying to sell your abilities to write things. The entire point of OpenAI as a company is to sell its LLM.
Why would that matter?
I could totally start a website, maybe call it “New York Stories”, read every single NYT article and then (working off my own memory, not copy/pasting the text) write the same story again and publish it. That would not be copyright infringement. In fact the NYT themselves do it all the time, publishing things that were originally reported elsewhere.
If you write it down and sell it you 100% do violate copyright
Google be in trouble then
Do you have paying customers that ask you for movie scripts and song lyrics like OpenAi does? If so, the above would be flat out copyright infringement.
Pffffft, this guy can’t recite “Forgot about Dre” from memory…