Thomson Reuters wins AI training copyright case — what this does and doesn’t mean
Thomson Reuters has won most of its copyright case against defunct legal startup Ross Intelligence, who trained an AI model on headnotes from Westlaw. [Wired, archive; opinion, PDF, case docket] Th…
You're viewing a single thread.
Being an Internet Old, I do worry about rulings that could narrow the scope of fair use, since good things do rely upon it, and there's always the risk of a sowing/reaping thing, but I won't pretend to have actual legal expertise.
With the wide-ranging theft that AI bros have perpetrated, AI corps' abuse of the research exception, AI's ability to directly compete with original work (Exhibit A) and the myriad other things autoplag has unleashed on us, I suspect we're gonna see a significant weakening of fair use.
Giving a concrete prediction, the research exception's gonna be at high risk of being repealed. OpenAI et al crossed the Rubicon when they abused it to launder their "research data" into making their autoplags - any future research case will have to contend with allegations of being a for-profit operation in disguise.
On a more cultural front, if BlueSky's partnering with ROOST and the shitshow it kicked off is any indication, any use (if not mention) of AI is gonna lead to immediate accusations of theft. Additionally, to pull up an old comment of mine, FOSS licenses are likely gonna dive in popularity as people come to view any form of open-source as asking for AI bros to steal your code.
