
See this thread and several related ones: https://twitter.com/jess_miers/status/1732449736232497665
The text is:
"(5) GENERATIVE ARTIFICIAL INTELLIGENCE. The term 'generative artificial intelligence' means an artificial intelligence system that is capable of generating novel text, video, images, audio, and other media based on prompts or other forms of data provided by a person."
As a matter of practical law, if this were to pass, would it be applied to at least one of YouTube or Facebook's curation algorithms, in a way that held without loss of generality either Meta or Google liable for user content, as some people are warning could be the case?
Resolves YES or NO if this actually happens and a verdict is reached (also see house rules for early resolution). Will resolve on any case against a tech company that settles the legal question in either direction.
Resolves to fair market price (or N/A if necessary) if point becomes moot, or the deadline occurs and no case is pending.
There have been suggestions this could even apply to spellcheckers or in other even crazier situations - if anyone thinks that should trade >10% I will consider opening those markets as well.
People are also trading
Betting yes because the price is so low.
My confidence would be a lot higher if the question were 'will there be a court case about this?'
Collages are already held to be 'novel' works, I think there is a conversation about whether a collection of existing media is 'novel' on the basis of how it is curated.
I suspect the ruling will hinge upon whether the judge leans towards the 'well we all know what the lawmakers meant' or 'lawmakers should write good laws, fixing their mistakes isn't my job' side of the fence...
generating novel
On a first glance you'd think that NO, different combinations of existing videos do not constitute novel media content.
Then you have the whole LLMs interpolating, and not extrapolating debate, that they're not generalizing but learning statistical correlations of existing media (mainly text) in novel ways, similar to how curation algorithms work.
This is going to come down to whether LLMs generalize and produce novel content, which will have implications with copyright law.