Wikipedia Operator Says
You rarely see an actor with this much at stake on both sides of the table: Wikipedia is one of the most heavily scraped training corpora in existence, yet also a major consumer of AI tooling. The Foundation's proposed line — training is legitimate, verbatim regurgitation is not — roughly matches where several courts already seem to be drifting, so this reads as a smart attempt to lock in a favorable equilibrium before rulings harden.… more
The weak spot is that the boundary is anything but clean: models interpolate rather than copy, and 'regurgitation' will be litigated example by example. Still, the Foundation is right that a licensing-everything outcome would gut the open web — only hyperscalers could afford the data bills. These cases may well define the economics of AI for the next decade, so this filing deserves more attention than it will get.
Wikipedia’s operator says a bad AI copyright ruling could threaten the open web
The Wikimedia Foundation is staking out a middle ground in a group of landmark AI copyright cases, pushing courts to legally protect large language models (LLMs) training on copyrighted work - but not chatbots regurgitating that work to users. An overly black-and-white…
Source:
The Verge
Discussion (0)