Supreme Court's Cox Ruling Reshapes Copyright Fight, With AI Cases Next in Line
The high court's 9-0 Cox v. Sony ruling hands AI developers a ready-made defense against claims that their models are built for infringement.
WASHINGTON, July 15, 2026 – The Supreme Court's unanimous decision shielding internet service providers from liability for their subscribers' copyright infringement has established a clean legal test that will likely define the wave of artificial intelligence copyright litigation now working through the courts, the Electronic Frontier Foundation's legal director said Wednesday.
Speaking on Broadband Breakfast Live Online, Corynne McSherry said the March ruling in Cox v. Sony reaffirmed decades-old precedent that general purpose tools cannot be held responsible for every misuse. Liability attaches only when a company induces infringement or builds a product tailored for it.
Broadband Breakfast
"We now have a nice, clean, clear test," McSherry said. "It's predictable. You really want that, especially if you're a business who's just trying to operate, but also if you're a user who's relying on that business."
A unanimous court, and why it landed that way
Sony sued Cox Communications after sending hundreds of thousands of infringement notices, arguing the ISP failed to terminate repeat infringers. A jury returned a $1 billion verdict on a theory of contributory liability, upheld by the Fourth Circuit Court of Appeals, before the Supreme Court reversed 9-0.
McSherry, whose organization filed an amicus brief on behalf of Cox, said the justices grasped what was at stake for ordinary subscribers. "Everyone needs internet access these days to do everything, to go to school, to access government services, to access the library," she said. "It's sort of its basic infrastructure for people."
During oral argument, justices pressed Sony's counsel on whether ISPs should cut off universities and coffee shops over one user's conduct. "Sony's lawyer really didn't have a good answer for that," McSherry said.
Drew Clark, CEO of Broadband Breakfast, noted the rare unanimity. "This is a 9 to 0 decision. That doesn't happen that often," he said. "Did you expect that, and why do you think it ends up being unanimous?"
McSherry credited the unanimity on timing.
"When it comes to technology, that is a very good thing, because what you want is for basically a technology to have a chance to gel a little bit," she said. "You've got clerks and the justices themselves who understand the potential implications of ruling the other way."
Clark drew a line back to the statutes themselves. "The 1996 telecom act was followed by the 1998 digital millennium copyright," he said. "Super productive period for the foundational laws affecting the internet and telecommunications."
From Betamax to the LLM lawsuits
The reasoning in the Cox case traces to the 1984 Betamax case, also involving Sony, where Hollywood argued the VCR would destroy the film industry.
The court declined to rewrite rules about copyright because of a new technology, leaving that to Congress.
That restraint might be necessary now, said McSherry, with roughly 100 copyright suits filed against AI developers.
The plaintiffs in those cases argue that works were copied for training and that large language models function as infringement engines producing competing works. The second claim, McSherry said, is a secondary liability theory that Cox directly undercuts. "LLMs do all kinds of things. They are not infringement machines," she said.
Clark pushed back on the idea that users bear the blame. "That just seems like it strains common sense," he said, "to say that the user is the one infringing rather than the LLM that's got that repository of data."
But McSherry held her ground. "People use all kinds of tools to make copies of things," she said. "But that doesn't mean the person that provided that tool is on the hook for it."
Sections 1201 and 1202 of the DMCA, governing anti-circumvention and copyright management information, are "alive and kicking," she added, and now appear in AI complaints.
Privacy and California law
On privacy, McSherry warned that AI supercharges existing surveillance and urged a comprehensive federal law rather than AI-specific provisions, pointing to state action as fertile ground. California, she noted, retains net neutrality protections.
"We fight for the user," she said.
