A coalition of publishers, news media, musicians, photographers and film producers has rejected a proposal issued earlier this week by the U.K. government to allow artificial intelligence companies to train their large language models on their content if they don’t opt out.

Since the rise of generative AI, tech companies have found themselves in legal tangles with the artists who create the work on which AI products are trained. While there’s no easy fix, the U.K. has mulled over an opt-out model in which the creatives must take the initiative to ask companies to keep their hands off their work. This has largely been dismissed by the creatives, who prefer an “opt-in” model.

This week the Labor government introduced a set of proposals in an effort to placate the artists who feel their copyright is being infringed while still giving AI developers “wide access to material to train world-leading models.” The proposal was based on an opt-out model that would create a loophole in which AI companies should skirt copyright laws.

Today, the Creative Rights in AI Coalition said the onus should not be on the owner to opt out but on the AI to seek permission and ink a licensing deal. The coalition, a list of heavyweights, includes the Guardian, Financial Times, Telegraph, the Daily Mail Group and Getty Images, and also the Motion Picture Association, the Society of Authors, the British Phonographic Industry and the Independent Society of Musicians.

“Rights holders do not support the new exception to copyright proposed,” the group said in a statement, which was shared with The Guardian. “In fact, rights holders consider that the priority should be to ensure that current copyright laws are respected and enforceable. The only way to guarantee creative control and spur a dynamic licensing — and generative AI — market is for the onus to be on generative AI developers to seek permission and engage with rights holders to agree licenses.”

In a recent poll in the U.K., 72% of respondents said AI companies should have to pay royalties to the creators of content that has been used to train AI, while 80% said those same companies should make it clear just what content they’ve used.

Across the pond, various deals have been struck with AI companies and publishers, while OpenAI, Microsoft Corp., Google LLC and Meta Platforms Inc. have found themselves facing lawsuits for allegedly breaching copyright laws.

Image: Freepik