CEO says ‘control and compensation’ equally important in blanket licence model | AI companies would pay fees to use works under scheme that authors could opt into.
AI companies could be offered a ‘blanket licence’ to legally use content from books to train their chatbots, under a new proposal the Authors Guild is investigating.
The largest professional organisation for published writers in the US is exploring an opt-in licensing model as a way to compensate authors when their works are used for generative AI—and give them a level of control.
Its chief executive, Mary Rasenberger, told The Hollywood Reporter last week: “We have to be proactive because generative AI is here to stay.”
“Our position is that there’s nothing wrong with the tech, but it has to be legal and licensed,” she said.
The scheme, which could be called the ‘Author’s Registry’, would involve AI companies paying fees to use books to train their models as well as for outputs that reference content.
Authors Guild members could choose whether or not to have their works included in the licensing deal. Rasenberger said that among 2,400 members surveyed, around 38% indicated they would be interested in taking part.
Possible restrictions would include chatbots being prevented from producing material ‘in the style’ of certain authors and creating summaries of books.
The issue of control was “as important as the compensation”, Rasenberger stressed. She said although the project wouldn’t get off the ground this year, it had funding and was backed by best-selling US author James Patterson.
Copyright court battles
In September last year, Authors Guild members including John Grisham and George RR Martin sued ChatGPT owner OpenAI for allegedly copying their works “without permission or consideration” and feeding them into their algorithms.
The Authors Guild was exploring the licensing deal partly because AI companies were expected to argue in court that they had “no choice but to use materials sourced from shadow library sites to train chatbots since it wouldn’t be practical to individually negotiate licences with thousands of authors”, The Hollywood Reporter said.
Media firms CNN, Fox and Time are also having discussions around licensing deals with OpenAI, according to Bloomberg News.
The ChatGPT creator is reportedly looking to license articles from CNN to train its large language models (LLMs) and feature CNN’s content in its products.
As well as the lawsuits filed by authors, OpenAI is facing a copyright infringement suit from The New York Times which says it used “millions” of its articles to train chatbots.
The Microsoft-backed tech company responded in a blog post last week (January 8) that it was “disappointed” by NYT’s lawsuit which it considered to be “without merit”.
Did you enjoy reading this story? Sign up to our free daily newsletters and get stories sent like this straight to your inbox
