Signing licensing deals with artificial companies can help protect good journalism, argues Thomson Reuters CEO Steve Hasker.
Katie Prescott of The Times of London writes, “As boss of the enormous content company, Hasker has overseen a number of licensing deals with AI businesses in order to sell its information to power their large language models. These have included a tie-up with Meta Platforms. It uses models from OpenAI to underpin its legal AI assistant called CoCounsel.
“‘Reuters has a historical news file as a reference data set for these models. Our news is free of bias. It is independent. It’s fact based. It is triple checked and verified. We don’t offer opinions. We’re not leaning left or leaning right or spinning a narrative for the entertainment of our listeners, viewers or readers. It really is the underlying facts associated with the news. So it’s a particularly powerful source for these models to be trained on or built on,’ Hasker said.
“Many content businesses were taken by surprise by the advent of large language models and the realisation that their intellectual property may have been used to train them, without their knowledge or consent. Hasker said: ‘Licensing deals clear the air in terms of the legalities of that use. I think there’s definitely an element of that.'”
Read more here.
