Posted on August 22, 2023 at 2:00 PM by Sadye Scott-Hainchek
The Atlantic has identified thousands of authors whose pirated works were used to train artificial intelligence tools.
Atlantic subscribers can read the story in full; the Guardian provided a glimpse at the scope of the investigation and its revelations.
Companies including Meta and Bloomberg used a dataset known as Books3, which appears to contain more than 170,000 identifiable and pirated books, to train their generative AI tools.
Among the authors of those works: Margaret Atwood, Junot Díaz, Jennifer Egan, Elena Ferrante, Jonathan Franzen, bell hooks, Stephen King, Jon Krakauer, Haruki Murakami, Michael Pollan, George Saunders, Zadie Smith, and Rebecca Solnit.
The Guardian reports that Bloomberg acknowledged using Books3 and said it won't use it to train any more AI tools, but Meta didn't respond to comments.
The Atlantic's reporting comes after the writers Sarah Silverman, Richard Kadrey, and Christopher Golden filed a lawsuit claiming their copyrighted works were used to train Meta's AI.
Related posts
Categories: Today in Books