Login
iscover
Share this article on:
Powered by
MOMENTUM
MEDIA
Powered by MOMENTUMMEDIA
The head of the US Copyright Office has been fired less than a day after she refused to allow Elon Musk to mine copyrighted data to train his AI models.
News of the termination of Shira Perlmutter was first reported by Politico and CBS News, and it appears to have been confirmed by Committee for House Administration representative and Democrat Joe Morelle.
“Donald Trump’s termination of the register of copyrights, Shira Perlmutter, is a brazen, unprecedented power grab with no legal basis,” said Morelle.
“It is surely no coincidence he acted less than a day after she refused to rubber-stamp Elon Musk’s efforts to mine troves of copyrighted works to train AI models.”
Morelle also linked a pre-publication version of a US Copyright Office report that defines the terms through which copyrighted works can be used for AI training, leading to the conclusion that the termination is related to Musk’s denied request to train AI on copyrighted content.
“When a model is deployed for purposes such as analysis or research – the types of uses that are critical to international competitiveness – the outputs are unlikely to substitute for expressive works used in training,” the report said.
“But making commercial use of vast troves of copyrighted works to produce expressive content that competes with them in existing markets, especially where this is accomplished through illegal access, goes beyond established fair use boundaries.“
Trump also seemingly shared the news of the termination on his social media platform Truth Social, where he reposted or “ReTruthed” a post that strangely appears to be criticising the termination, and links to the original CBS News article.
“Now tech bros are going to attempt to steal creators’ copyrights for AI profits,” said attorney Mike Davis.
“This is 100 per cent unacceptable.”
It’s not just Musk who is interested in copyrighted works for AI training.
OpenAI has previously said it cannot train large language models (LLMs) without copyrighted data.
“Because copyright today covers virtually every sort of human expression – including blogposts, photographs, forum posts, scraps of software code, and government documents – it would be impossible to train today’s leading AI models without using copyrighted materials,” it said in a submission to the House of Lords communications and digital select committee in January.
Be the first to hear the latest developments in the cyber industry.
