“Foreign corporations that enjoy ‘concerted state support’ have ‘unfettered’ access to ‘copyrighted data’ that will ‘improve’ their AI models. But under the district court’s opinion, ‘American companies are left without fair use access’ and the ‘race for AI is effectively over.’”

https://t.co/OBr6iH3VmK

ROSS was so ahead of the curve, it still shocks me. and I was there. believing that 'RL will enable machines to learn on their own' and building toward autonomous learning in 2017 is absolutely bonkers. also insane alpha.

“The resulting opinion [of @ThomsonReuters v. @ROSSIntel] could shape how courts approach the core fair-use inquiries: what counts as “transformative,” and how the possibility of licensing training data should factor into the market-harm analysis.”

“The odds of a significant ruling are pretty high. Several interlocutory appeals are already in motion, and none looms larger than Thomson Reuters v. Ross Intelligence, now before the Third Circuit.” 🔗

Litigation Revives The Strange Tragic Tale Of John West, The Founder Of Modern Legal Research via @bobambrogi https://t.co/GYTQI3vWDb

Germany becomes the latest jurisdiction to hold that AI training on copyrighted works is lawful. In Kneschke v. LAION the court confirmed that dataset creation falls under the text and data mining exception, and that subsequent commercial use does not change the analysis.

Plenty of mainstream outlets have covered the copyright lawsuits against American AI labs. Why haven’t any of them noted the Chinese models they compete with use the same libraries and sources, but face no legal scrutiny?

FOLLOW UP: “How does this relate to the ROSS Intelligence v. Thomson Reuters appeal?”

ANSWER: “The lawsuits are therefore aimed at US companies not to stop copying, but to extract rents, slow domestic competitors…this turns copyright enforcement into an industrial policy that penalizes reachable US firms while unreachable foreign actors operate freely.”

Great overview of the @ROSSIntel v. @ThomsonReuters appeal: The Third Circuit need not choose between originality and fair use. It can hold that Westlaw’s headnotes are not entitled to copyright protection and, in any event, that ROSS’s use would be fair even if they were.

"On one hand, the court could alter the trajectory of the more than 60 copyright infringement cases against AI companies winding their way through US district courts by squarely addressing the intersection of fair use with artificial intelligence. On the other, it could skirt the

"The deeper truth is that AI is less a rival to human creativity than a multiplier of it—allowing people to spend more time on what technology can’t replicate: judgment, taste, voice, imagination." https://t.co/IFx0Ot8fF6

https://t.co/KCaIRLjvUR

something everyone can agree on: “the lower court’s holding establishes key precedent”

The decision in ROSS v. Thomson Reuters will be binding on all the other AI cases in the country as the first appellate decision on fair use and training an AI model. As goes ROSS, so goes AI innovation in the United States of America.

.@Disney and @OpenAI have a 3 yr deal! Details to follow. Key is 200 D characters now open for Sora users while OAI will block others from *outputs* & protect actor voices/faces. Training *inputs* not addressed. Savvy OAI suit management & D brand boosting in new market. Win/win?

Great that @ROSSIntel v. @ThomsonReuters is featured in @Axios' coverage of the OpenAI-Disney partnership today. AI licensing deals are evolving to focus on outputs, not inputs (training), which preserves fair use in the US as a foundational pillar of innovation. Tide is turning!

What is fair use?

“If everyone has access to key legal documents, they can build innovative new products with them, especially those that are based on artificial intelligence and machine learning—which require access to a lot of data to work at all.” - Jacob Heller, Casetext cofounder

“Questions about how copyright law should be applied to Al are already playing out in the courts. That's where this issue will be decided.” - @DavidSacks 👀