More than 10,000 members of the Authors Guild, America’s largest organization for writers, recently co-signed anurging AI leaders at tech companies like Open AI and Meta to stop using their work without permission or compensation. “Millions of copyrighted books, articles, essays and poetry provide ‘food’ for AI systems, endless meals for which there has been no bill,” says the letter, featuring the signatures of industry bold faces like Margaret Atwood, Jennifer Egan and Dan Brown.
When you look at the history of new technology, it doesn’t get stopped. If it can be done, somebody will do it. There’s this feeling of inevitability, like we’re trying to stop a boulder from rolling down the hill. I think that’s why the message is: “If we can’t stop it, let’s find a way to control it.” If you’re going to feed our books and our material into this computer, then we want consent and compensation.
How can you tell that the generative AI tools are being trained on the copyrighted work of published authors and not, say, Reddit threads and Wikipedia posts?