New York Times Co. v. Sullivan and New York Times Co. v. United States (Pentagon Papers) are two famous examples of a great flowering of First Amendment jurisprudence during the middle of the twentieth century. The philosopher Alexander Meiklejohn declared Sullivan to be “an occasion for dancing in the streets.” Sullivan recognized that “the central meaning of the First Amendment” was that the state could not punish criticism of public officials made without malice either directly through the criminal law or indirectly through civil damages for defamation. Pentagon Papers reaffirmed the central First Amendment principle against prior restraints; Justice Stewart’s concurring opinion added that the government could not suppress disclosure of sensitive information unless the disclosure would “surely result in direct, immediate, and irreparable damage to our Nation or its people.” Together these two decisions celebrated the crucial role of the press in a democratic society, and stood for the principle that the circulation of public discourse is crucial to democratic legitimacy. Half a century later, the impact of these two decisions has been weakened by significant changes in the practices and technologies of free expression, changes that concern a revolution in the infrastructure of free expression. That infrastructure, largely held in private hands, is the central battleground over free speech in the digital era.
Government practices have also changed in the past fifty years. To be sure, governments still regulate speech through fines, criminal penalties, and injunctions; they still engage in predigital practices of surveillance. But new techniques have supplemented traditional modes of control over speech and traditional modes of surveillance. Like speech itself, the regulation and surveillance of speech require an infrastructure. Increasingly, speech regulation and surveillance are technologically imposed and involve cooperation between governments and the private entities that control the infrastructure of free expression.
Thus, a significant feature of the early twenty-first century is that the infrastructure of free expression increasingly is merging with the infrastructure of speech regulation and the infrastructure of public and private surveillance. The technologies and associated institutions and practices that people rely on to communicate with each other are the same technologies and associated institutions and practices that governments employ for speech regulation and surveillance.
Consider a mid-twentieth-century newspaper like the petitioner in New York Times Co. v. Sullivan. To reach its audience, the Times depended on an infrastructure of technologies and institutions: printing presses, labor unions, delivery trucks, newsstands, and advertisers. These features of the Times’s business may have been regulated by the government in various ways — in trucking regulations, labor law, and so on. But for the most part the government’s capacities for control and surveillance of speech were not built into the very technologies and practices that the Times used to communicate with its audience. The government did not have a long-distance switch that allowed it silently and inexpensively to control the Times’s printing presses or prevent certain articles from appearing in its pages. The government did not require that members of labor unions operating the Times’s printing presses wear hidden microphones and cameras so that the government could learn about any potentially subversive or infringing materials. That is why it was necessary for the government to seek an injunction in the Pentagon Papers case. Of course, the government did control the public streets. Arguably it could have created roadblocks throughout New York City to search for and stop the Times’s delivery trucks, but this would have been highly visible, logistically difficult, and costly in terms of legitimacy.
The digital era is different. Governments can target for control or surveillance many different aspects of the digital infrastructure that people use to communicate: telecommunications and broadband companies, web-hosting services, domain name registrars, search engines, social media platforms, payment systems, and advertisers. The very forces that have democratized and decentralized the production and transmission of information in the digital era have also led to new techniques and tools of speech regulation and surveillance that use the same infrastructure. These tools of regulation and surveillance often work automatically and in the background; they may harness the cooperation or coercion of private owners of infrastructure to achieve the government’s regulatory goals. Low salience and use of private parties can help governments preserve legitimacy even as their policies block, limit, or spy on expression. This is the big story about the freedoms of speech, press, and association in the digital age.
Traditional or “old-school” techniques of speech regulation have generally employed criminal penalties, civil damages, and injunctions to regulate individual speakers and publishers. The landmark decisions in Sullivan and Pentagon Papers responded to old-school speech regulation: in both cases, the state had used penalties and injunctions directed at speakers and publishers in order to control and discipline their speech.
These methods have hardly disappeared in the twenty-first century. But now they are joined by “new-school” techniques of speech regulation. The latter regulate speech through control over digital networks and auxiliary services like search engines, payment systems, and advertisers; instead of focusing directly on publishers and speakers, they are aimed at the owners of digital infrastructure.
These new-school techniques have three characteristic features that often operate together. None of these features is entirely new. Each has counterparts or precedents in the predigital world, but each has been reshaped to fit the demands of a new technological environment.
The first feature is collateral censorship, in which the state regulates party A in order to control speaker B. The digital age enables a vast number of people to communicate widely across the country and around the world. Because there are so many speakers, who are often anonymous, difficult to co-opt, or otherwise beyond the government’s effective control, the state aims at Internet intermediaries and other owners of digital infrastructure — threatening liability to induce them to block, limit, or censor speech by other parties.
Second, and relatedly, public/private cooperation and co-optation are hallmarks of new-school speech regulation. To the extent that the government does not own the infrastructure of free expression, it needs to coerce or co-opt private owners to assist in speech regulation and surveillance — to help the state identify speakers and sites that the government seeks to watch, regulate, or shut down. To this end, the government may offer a combination of carrots and sticks, including legal immunity for assisting the government’s efforts at surveillance and control. Owners of private infrastructure, hoping to reduce legal uncertainty and to ensure an uncomplicated business environment, often have incentives to be helpful even without direct government threats.
Third, governments have devised new forms of digital prior restraint. Many new-school techniques of speech regulation have effects similar to prior restraints, even though they may not involve traditional licensing schemes or judicial injunctions. In addition, prior restraints are especially important to the government’s expansive surveillance practices in the National Surveillance State. As I explain in Part III, prior restraints directed at owners of private infrastructure are now ubiquitous in the United States; gag orders have become fully normalized and bureaucratized elements of digital surveillance, as routine as they are invisible.