US urges G20 countries to allow AI training on creators’ work

The U.S. has called on G20 nations to establish guidelines regarding how AI companies utilize copyrighted content for training their systems, as numerous American technology firms face litigation over this practice. Tech giants such as Anthropic, OpenAI, Google, and Meta are embroiled in lawsuits filed by authors, publishers, media outlets, and other creators who claim these firms have improperly used their work to develop their AI products.

These companies train their AI models using extensive datasets, which often include books, music, and art. They argue that employing this material to build their AI tools falls under the legal umbrella of fair use. During a meeting in North Carolina focused on technology, U.S. Commerce Secretary Howard Lutnick told G20 officials that they should permit AI companies to train their models on creators' work while simultaneously finding a way to safeguard artists.

He emphasized the need for frameworks that protect creators and innovators without hindering the development of future innovators. The U.S. Justice Department also filed a brief on Wednesday backing OpenAI in its legal battle with the New York Times and several other newspapers over the use of their work to train the language models behind ChatGPT.

According to the brief, AI training generally qualifies as fair use of copyrighted material. Nvidia CEO Jensen Huang addressed the G20 officials before Lutnick's speech, urging them to avoid regulations that emphasize potential theoretical harms and instead focus on developing rules addressing real-world problems associated with AI.

Huang's remarks echoed those of the Trump administration and other American tech leaders who spoke at the event on Tuesday. Government regulations may negatively impact the industry's profitability if they delay the release of new models or force companies to alter their products' performance to address security concerns. Anthropic co-founder Tom Brown, Palantir's Alex Karp, and OpenAI's Sam Altman are scheduled to speak at the event in the afternoon.