It seems that the US Government (NSA, CISA, and FBI) is crying foul over something they themselves are defending. They’re suggesting that it’s ok for a US company to train on copyrighted materials, but a week later they’re suggesting that non-US companies doing the same is a national security threat.
Historically, civilizations have built upon the works of other civilizations, just look at most mythological/religious/ancient texts. One could say that Exodus was distilled from the Laws of Hammurabi.
This may be the first time it’s happening and observable in real time, but it’s not like the frontier companies didn’t also rip off the rest of the world. There are countless lawsuits about the intellectual property rights of authors, and other published works. No one is questioning whether the companies did it, they’re asking to be fairly compensated for their work. The accusation that the Chinese companies are using burner accounts and circumventing access controls is not much different from what the frontier companies have been accused of themselves. At least the frontier companies are being paid for accessing their content, can’t say the same for the authors who had to sue.
The lawsuits are so many that there’s even a website https://ailawsuittracker.com/copyright/ that tracks AI copyright lawsuits.
I’m not sure I agree with the allied nations part of this statement.
“Addressing industrial-scale distillation merits a coordinated response across the AI ecosystem, including effective information-sharing, spanning the U.S. Government, private industry, and allied nations”
Where does Sarvam fit in? They’re saying training was conducted in India, but where was the data actually collected? Did they not do any distillation? Or, could it be that since no one I know has heard of these models, they aren’t posing a threat to the US based companies? tbh I just found out about these models as I was doing research for this post. I’ll have to try them out.
Jakub Pachocki makes the case for a coordinated response across the AI ecosystem, which I support.
We’re on the cusp of making huge leaps in technology based on the advancements in AI. I really hope we don’t go backwards by gatekeeping the AI ecosystem.