Sen. Josh Hawley (R-MO) accused Meta of “willfully pirating” vast amounts of copyrighted material to train its artificial intelligence models during a Senate Judiciary Crime & Terrorism Subcommittee hearing on Wednesday.
Chairing the session, Hawley highlighted what he described as an intentional and illegal effort by Meta to ingest protected works despite repeated internal warnings from employees.
“They knew exactly what they were doing,” Hawley said. “Employee after employee warned management that this was illegal. One even shared an article about the chances of getting arrested for torrenting. Another warned that downloading from illegal sources would expose Meta to legal risks.”
“Did Meta’s leadership listen? No. They bulldozed straight ahead,” he continued.
The hearing featured testimony from bestselling author David Baldacci, legal scholars, and AI experts, all raising alarm over tech companies allegedly exploiting copyrighted material without permission. Baldacci expressed outrage over discovering his entire library was accessible to an AI system within seconds. “It felt like I’d been robbed of everything I’ve worked for in my adult life,” he said.
Hawley further alleged that Meta CEO Mark Zuckerberg personally approved the use of pirated content and that executives made efforts to conceal the practice, even using non-company servers.
“Meta trained its AI model to lie to users about what data it was trained on,” Hawley claimed, calling the company’s behavior “not aggressive competition, but criminal conduct.”
He added that Meta’s alleged infringement is part of a broader, systemic issue across the AI industry, accusing large tech firms of routinely ignoring copyright law for the sake of faster model development.
According to a press release from Hawley’s office, AI firms may have used over 200 terabytes of copyrighted content without proper licensing.
The debate over fair use in AI training has become a flashpoint in Washington. White House AI Advisor David Sacks previously argued that without some level of fair use, U.S. AI development would lag behind countries like China. “China is going to train on all the data regardless,” Sacks said in June, “so without fair use, the U.S. would lose the AI race.”
