Congress Wants Tech Companies to Pay Up for AI Training Data
Lawmakers call for OpenAI and other tech companies to pay for licensing news articles and data used to train AI algorithms.
Growing Consensus in Congress
The question of whether AI companies should pay for the training data they use in their generative AI systems is a controversial topic in Silicon Valley. However, in Washington, DC, lawmakers at a Senate hearing on AI's impact on journalism agreed that tech giants like OpenAI should compensate media outlets for using their work.
Richard Blumenthal, the Democrat who chairs the Judiciary Subcommittee on Privacy, Technology, and the Law, stated, "It's not only morally right. It's legally required." Republican Senator Josh Hawley also expressed support, emphasizing the need to prevent tech companies from freely using valuable data without compensation.
Media Industry Support for Licensing
During the hearing, media industry leaders spoke in favor of licensing news articles and data for AI projects. The CEO of the National Association of Broadcasters, the CEO of the News Media Alliance, and the CEO of Condé Nast all highlighted the negative impact of AI companies using their content without compensation. They argued that such actions undermine the quality of content and infringe copyright laws.
Roger Lynch, the CEO of Condé Nast, urged lawmakers to clarify that using journalistic content without proper licensing agreements does not fall under fair use. The media industry leaders emphasized the necessity of licensing agreements to ensure a fair market.
Disagreements and Potential Impacts
While there was consensus among lawmakers and media industry insiders at the hearing, not everyone agrees that mandatory licensing is the best solution. Some argue that the practicality and economic burden of licensing all training data is unrealistic.
Sarah Kreps, director of the Tech Policy Institute at Cornell University, expressed concerns about the feasibility of licensing requirements and its potential to favor big tech companies over startups. Northwestern computational journalism professor Nick Diakopoulos also expressed skepticism about a mandatory licensing regime.
Legislation is being considered to address data licensing, including bills that authorize news outlets to negotiate licensing arrangements and a bipartisan framework that proposes a licensing regime overseen by an independent body. The exact approach and outcome are yet to be determined.