The world’s top AI models can be prompted to generate near-verbatim copies of bestselling novels, raising fresh questions about the industry’s claim that its systems do not store copyrighted works. A series of recent studies has shown that large language models from OpenAI, Google, Meta, Anthropic, and xAI memorize far more of their training data than previously thought.
Source: AIs can generate near-verbatim copies of novels from training data



The Australian government should ignore claims from artificial intelligence giants that local copyright laws prevent them from investing heavily in data centers, say media and entertainment industry leaders. “Where is the proof that issuing a copyright exemption or law change would bring massive economic benefit to Australians and not just major US tech companies?” said head of the Australian Recording Industry Association, Annabelle Herd.
“It’s entirely possible that in two years we’ll have clarity that’s probably good enough in terms of a set of best practices,” estimated Jessica Litman. But she warned it could be longer, “because these cases are so huge, they are happening in dribs and drabs… so it can take a while to establish a trend.” She said it could be as many as 10 more years until these lawsuits are resolved.


