Prof. Goldstein on Cox v. Sony (Excerpt from His Treatise)
Prof. Paul Goldstein (Stanford Law) kindly has allowed me to share this update to his treatise (Goldstein on Copyright, Third Edition) regarding the Supreme Court’s Cox v. Sony decision. (My initial comments on the ruling are here).
* * *
In Cox Communications, Inc. v. Sony Music Entertainment, an ill-considered 2026 decision that will serve the interests neither of copyright owners nor of copyright users—nor, for that matter, society’s interest in a balanced copyright system—the United States Supreme Court took an axe to contributory liability, ruling that an internet service provider serving approximately six million subscribers was not contributorily liable for its subscribers’ copyright infringements merely because it had provided “a service to the general public with knowledge that it will be used by some to infringe copyrights.” According to the Court’s opinion, written by Justice Clarence Thomas, “[t]he provider of a service is contributorily liable for a user’s infringement if it intended its service to be used for infringement.” To establish such intent, “a copyright owner must show one of two things. First, it can show that a party affirmatively ‘induc[ed]’ the infringement. Or, second, it can show that the party sold a service tailored to infringement.”
Image produced by ChatGPT March 2026
“Inducement” has an established place in copyright jurisprudence. But the concept, “tailored to infringement,” appears nowhere in the Copyright Act or in copyright case law. The late Justice Ruth Ginsburg’s concurring opinion in Metro-Goldwyn-Mayer Studios Inc. v Grokster Ltd., to which Justice Thomas ascribed the term, in fact nowhere included those words. Justice Ginsburg did employ the terms “substantial” or “commerciallv significant” noninfringing uses, referring to the concept that had become a central determinant for withholding contributory liability for providing goods or services since the Court first adopted it in Sony Corp. of America v. Universal City Studios, Inc. Justice Thomas’s treatment of the two terms as equivalent—“[a] service is tailored to infringement if it is ‘not capable of “substantial or commercially significant”’ non-infringing uses—overlooks not only that the first term is materially narrower than the second, but also that Sony borrowed it from the patent statute with the distinct object of ensuring that consumers not be required to pay monopoly tribute for otherwise unprotected goods or equipment. If selling salt tablets that can be used in a patented salt dispenser should not be actionable because the tablets have a substantial noninfringing use, so sales of video recording devices employed to copy copyrighted works should not be actionable if the devices possessed one or more substantial noninfringing uses, such as time-shifting recorded programs for later viewing.
By assimilating “substantial noninfringing use,” an objective, competition policy-based measure of liability, into a necessarily subjective intent-based measure, Cox seems likely to destabilize entrepreneurial planning for the introduction of new devices and services for exploiting copyrighted works. To be sure, service providers may see in the opinion a license to deliver services to subscribers–even though they know that many and even most of the uses may infringe–so long as they don’t promote or specifically design their services for infringement. In fact, if the Court believed that its narrowing of contributory liability to inducement by words or by device or service design, would enhance marketplace certainty, it might not have reflected carefully enough on the flimsy, indeed passive, evidence that it accepted in Grokster as evidence of inducement: that the defendants aimed “to satisfy a known source of demand for copyright infringement”; that defendants failed “to develop filtering tools” to “diminish the infringing activity using their software”; and “the commercial sense of their enterprise turns on high volume use.” Cox offers small solace to service providers if an act of omission such as failure to install filtering tools will subject them to liability.
The Court’s attempt in Cox to circumscribe contributory liability on the internet is understandable. The prospect of an aggregation of tens of thousands statutory damage awards—the award against Cox was $1 billion—can be daunting, indeed devastating, for even the most well-heeled service provider. The correct solution, however, would be for Congress to revise the Act’s statutory damages provision, and at least one bill has been introduced to that end. The prospect of secondary liability for generative AI platforms for the conduct of their users in prompting texts, sounds and images similar to copyrighted works may also have figured in the Court’s attempt at caution. But the question of secondary liability—indeed, of direct infringement liability—for these AI activities has not yet been directly addressed by the courts, or by the Congress, and for the Supreme Court to legislate on the question without a trial record or statutory text before it would be premature.
The terrain of secondary copyright liability generally has turned treacherous since the introduction in 1998 of the internet safe harbors, for the expectations of service providers and their subscribers have over the ensuing years been shaped less by secondary liability doctrines than by the mechanisms of the safe harbors. Lawsuits defining the boundaries of liability on the internet have been displaced in the vast number of potential conflicts by compliance with the safe harbors, including by automated means such as YouTube’s immensely successful Content ID system. If there has been litigation over service provider liability, it has been mainly to define the terms of the safe harbors and not to measure the reach of secondary liability doctrines. As proposed elsewhere in this treatise, the solution to this legal vacuum is for courts to populate it with the norms that business operations under the safe harbors have come to define as reasonable behavior among copyright owners and service providers alike. Sadly, the Court in Cox ignored this possibility. Indeed, as Justice Sonia Sotomayor wisely observed of the safe harbors in her concurrence in Cox, “[i]mportantly, Congress did not provide that ISPs could never be secondarily liable for copyright infringement. Instead, it struck a balance by creating incentives for ISPs to take reasonable steps to prevent copyright infringement on their networks, while also assuring ISPs that they do not need to take on the impossible task of responding to every instance of infringement on their networks….The majority’s new rule completely upends that balance and consigns the safe harbor provision to obsolescence.”