Do AI companies need to pay for the training data that powers their generative AI systems? The question has been hotly contested in Silicon Valley and in a wave of lawsuits against tech behemoths like Meta, Google and OpenAI. In Washington, DC, though, there seems to be a growing consensus that the tech giants need to cough up.
Today, at a Senate hearing on the impact of AI on journalism, lawmakers from both sides of the aisle agreed that OpenAI and others should pay media outlets to use their work in AI projects. “It’s not only morally right,” said Richard Blumenthal, the Democrat who chairs the Judiciary Subcommittee on Privacy, Technology and the Law, which held the hearing. “It is legally required.”
Josh Hawley, a Republican who worked with Blumenthal on AI legislation, agreed. “It shouldn’t be that just because the biggest companies in the world want to steal your data, they should be able to do it,” he said.
At today’s hearing, media industry leaders described how AI companies are harming their industry by using their work without compensation. National Association of Broadcasters CEO Curtis Leggett, News Media Alliance CEO Daniel Coffey and Condé Nast CEO Roger Lynch all spoke in favor of licensing. (WIRED is owned by Condé Nast.)
Coffey claimed that AI companies “deplete the quality content they feed on” and Lynch characterized the scraping of training data without permission as “stolen goods”. Both Coffey and Lynch also said they believe AI companies are infringing copyright under current law. Lynch urged lawmakers to clarify that the use of journalistic material without first brokering licensing agreements is not protected by fair use, a legal doctrine that allows copyright violations under certain conditions.
Common ground
The Senate hearings may have been adversarial, but today’s mood was largely favorable. Lawmakers and media industry insiders often compliment each other’s statements. “If Congress can clarify that using our content, or other publishers’ content, to train and output AI models is not fair use, then the free market will take care of the rest,” Lynch said at one point. . “That seems reasonable enough to me,” Holly replied.
Journalism professor Jeff Jarvis was the only dissenting voice at the hearing. He asserted that training on data obtained without payment was in fact fair use, and he spoke against mandatory licensing, arguing that it would harm the information ecosystem rather than protect it. “I must say that I am outraged to see publishers trading on the political capital earned through journalism to lobby for protectionist legislation,” he said, mocking his fellow speakers. (Jarvis was also subject to the only truly controversial question of the hearing from Republican Marsha Blackburn, who asked Jarvis about whether AI is biased against conservatives and the AI praising President Biden as evidence. Read the poem prepared by.)
Outside the committee room, there is less agreement that compulsory licensing is necessary. OpenAI and other AI companies have argued that licensing all training data is not feasible, and some independent AI experts agree.