The Two Paths to Contributory Liability
In Cox Communications v. Sony Music, the Supreme Court did not ask whether Cox could have done more to suppress infringement. It asked whether Cox's service and conduct showed the intent required for contributory liability. This chapter compares the two routes the majority recognized—active inducement and an infringement-tailored service—and then applies them to Cox. The result is a legal screen: general knowledge and imperfect prevention are not a third route simply because they feel culpable.
The distinction matters because a broadly useful service can be both foreseeable infrastructure for unlawful conduct and a legitimate product for millions of other users. If foreseeability alone established intent, the line between an ordinary tool and a service created to encourage infringement would disappear. The Court's precedents preserve that line by looking for evidence of purpose, not just evidence that misuse was predictable.
Inducement: when conduct reveals an objective
The first route is inducement. The majority described it as active encouragement of infringement through specific acts. The classic modern example is MGM Studios v. Grokster. The defendants distributed file-sharing software that had lawful uses, but the Court said a jury could find liability because the companies promoted and marketed the software as a tool for infringing copyright. The principal object of their business models was the use of the software to download copyrighted works.
That example adds something important to the earlier Betamax principle. A product does not escape liability merely because it can be used lawfully. In Sony Corp. of America v. Universal City Studios, the Betamax recorder could make unauthorized copies, but it also supported substantial lawful uses such as recording a television program for later personal viewing. In Grokster, by contrast, the provider's conduct supplied evidence of an affirmative purpose. The question was no longer just what the device could do. It was what the distributor was trying to make users do.
The Court's approach is therefore not a contest between “lawful” and “unlawful” products in the abstract. It compares capability with conduct. A general-purpose service may be capable of misuse without being promoted for that misuse. Conversely, a service may have lawful capabilities and still expose its provider to liability if advertising, instructions, business decisions, or other specific actions show an objective of encouraging infringement.
This is why inducement is more than a finding that infringement was common. A high level of unlawful use may raise questions about a provider's choices, but the majority's formulation looks for active encouragement. The evidence must connect the provider's own acts to the purpose of helping users infringe. The service's popularity among infringers is relevant context, not a substitute for that connection.
The distinction also prevents a provider from manufacturing distance through neutral language. If a service is publicly promoted as a way to obtain unauthorized works, or if users are instructed how to use it for infringement, a court need not accept the provider's description of the service as neutral. Purpose can be shown by what the provider actually does. The majority did not require a provider to confess an infringing objective; it required evidence from specific conduct rather than an inference based only on foreseeable misuse.
Cox did not fit that model on the record described by the Court. Sony did not present evidence that Cox marketed Internet access as a tool for downloading copyrighted music, instructed subscribers to infringe, or expressly promoted infringement. Cox's warnings, suspensions, and account terminations pointed in the opposite direction. They did not prove that Cox enforced its policy perfectly, but they weakened the claim that the company was affirmatively encouraging the acts.
The difference between Grokster and Cox can be stated as a question of direction. In Grokster, the provider's conduct directed users toward infringement even though lawful uses existed. In Cox, the provider supplied a connection that users could direct toward many purposes and responded to infringement reports with measures it said were intended to discourage further activity. The two services may both have been used for infringement, but the evidence of provider purpose ran in different directions.
The comparison also requires keeping two kinds of evidence apart. Evidence of provider conduct asks what the company said, marketed, instructed, or encouraged. Evidence of service capability asks what the product itself was built to do and whether meaningful lawful uses remained. The first can establish inducement even when the second points to a versatile product. The second can matter when the service is useful for little beyond infringement, even if the provider avoids explicit advertising. Cox presented neither combination: the record, as the majority described it, showed ordinary access and discouraging responses rather than an infringement-focused purpose.
That structure explains why the Court did not turn the case into a referendum on how much infringement is too much. A provider might receive a troubling volume of notices, yet the notices still must be connected to evidence about the provider's objective or the service's design. Conversely, a provider might avoid collecting detailed user information, but that absence would not protect it if its own promotion made an infringing purpose clear. The legal inquiry is demanding in both directions: it does not infer intent from misuse alone, and it does not allow a provider's chosen description to erase affirmative evidence of purpose.
Tailoring: when the service itself points toward infringement
The second route concerns the service's design and capabilities. The majority said a service is tailored to infringement when it is not capable of substantial or commercially significant noninfringing uses. The phrase does not mean that a provider is liable whenever a service has a harmful use, or that one lawful use automatically defeats a claim. It asks whether the service, considered as a product offered to users, is meaningfully useful outside the infringing purpose.
The Betamax recorder illustrates the negative answer. It could reproduce copyrighted television programs, but it could also record programs for later personal viewing. That lawful use was substantial enough that selling the recorder to the public did not itself amount to contributory infringement. The Court returned to the same principle in Cox: Internet access serves many purposes other than copyright infringement, and Cox did not tailor the connection to make copying easier.
The service-function question is different from the user-behavior question. Users may choose to use an ordinary connection for infringement, but their choice does not necessarily tell us what the provider's service was built to do. A connection that carries work, communication, education, commerce, entertainment, and other lawful traffic has a function broader than unauthorized copying. The provider may know that some users misuse it without having shaped the service around that misuse.
This is also where the network facts matter. Cox did not provide a library of selected music, a search system organized around unauthorized files, or instructions for obtaining particular recordings. It supplied Internet access. The Court said the service was not tailored to make copyright infringement easier. That conclusion followed from the ordinary character of the service and its substantial lawful uses, not from a finding that infringement was insignificant.
The boundary is visible in the Court's treatment of Grokster. Lawful capability did not prevent liability there because active promotion supplied the missing evidence of purpose. A service can therefore fall on different sides of the line depending on the interaction between design and conduct. A provider cannot rely on lawful uses if it simultaneously turns the service into an infringement-focused product. But a copyright owner also cannot establish tailoring merely by showing that a general-purpose service was used repeatedly for unlawful copying.
The majority's formulation should not be mistaken for a complete classification system for every online intermediary. The same company may operate a conduit, host user material, provide search or linking tools, or offer other functions. A rule about Cox's ordinary Internet access does not decide how a court should analyze a service that stores particular works, selects them for users, or organizes access around infringing material. The relevant service must be identified before its capabilities can be judged.
This is why “tailored” should be understood as a relationship between the service and the alleged use, not as a label attached to every product that can carry infringing material. Internet access can be essential to a lawful activity and also available to an infringing one. The Court's question was whether Cox changed that general function so that copyright infringement became the service's special purpose. It found no such change. A provider may still face a different analysis if it adds features, directions, or restrictions that make unauthorized copying the point of the offering.
The two routes also protect different kinds of reasoning. Inducement looks outward to communications and decisions that encourage users. Tailoring looks inward to the capabilities and ordinary uses of the service. They can overlap, but neither should be replaced by a general impression that the provider knew about infringement. Keeping them separate helps a court explain why a service is liable: because the provider promoted the wrongful use, because the product was effectively made for it, or because neither supported inference is present.
Why Cox falls outside both routes
Cox's response policy is useful for seeing why the two routes must remain separate. The company said that after a second notice it sent a warning; additional notices could lead to temporary interruption, a phone warning, or suspension; and after 13 notices an account could be terminated. Its contract prohibited subscribers from using the connection to post, copy, transmit, or disseminate infringing content. The Court characterized these actions as repeated discouragement rather than encouragement.
That conclusion does not mean the Court found Cox's response ideal. Sony emphasized that Cox terminated only 32 subscribers for infringement during the relevant period, despite terminating many more for nonpayment, and pointed to employee statements that Sony said reflected a preference for preserving subscription revenue. Cox answered that its warnings and suspensions ended 98 percent of identified infringement. Those facts could support an argument about how seriously Cox implemented an anti-infringement policy. They did not, without more, show that Cox promoted infringement or designed Internet access for it.
This is the crucial comparison. A warning is a provider action, but it is an action against infringement, not an invitation to commit it. A suspension changes a user's access, but it does not turn the underlying connection into a service tailored to copying. A limited number of terminations may support criticism of enforcement, but the majority held that insufficient prevention cannot be converted into inducement simply by relabeling it as knowledge plus inaction.
The Fourth Circuit had adopted precisely that conversion. It reasoned that supplying a product while knowing the recipient would use it to infringe was “exactly” the kind of culpable conduct that supported contributory infringement. The Supreme Court rejected the formulation because it went beyond the two forms recognized in Sony and Grokster. The Court also repeated an older warning: contributory liability cannot rest merely on a failure to take affirmative steps to prevent infringement.
The rule is demanding because it separates moral discomfort from a legal element. A provider may appear indifferent, may make commercially motivated decisions, or may fail to impose every possible consequence. Those facts can matter under another doctrine or statutory condition. But under the majority's contributory-liability framework, they do not replace proof that the provider intended its service to be used for infringement.
What would move a future case closer to one of the two routes? The evidence would need to add purpose or service-specific design. Promotion directed toward infringement would support inducement. Instructions that help users carry out unauthorized copying could do the same. A service with no substantial or commercially significant lawful use could support the tailored-service theory. Evidence that a provider specially shaped access around infringing activity would be more probative than evidence that ordinary users happened to misuse an otherwise broad service.
What remains insufficient by itself is equally important. General awareness that users may infringe does not establish inducement. Repeated notices do not automatically prove that the provider intended each act to succeed. Continued service, without evidence of promotion or tailoring, is not enough under the majority's rule. Nor does the existence of a response policy answer the merits question by itself. The evidence must be interpreted in relation to the service's function and the provider's purpose.
That is the doctrinal change Cox made to the Fourth Circuit's approach. It did not make knowledge irrelevant, and it did not bless services that invite or facilitate piracy. It refused to treat ordinary access plus knowledge and imperfect prevention as a third category of contributory liability. The next question is why the DMCA still imposes separate conditions on providers when substantive liability has not been established.