Executive Summary

• Some academics and pundits have proposed classifying aspects of the internet as “common carriers,” arguing this classification would provide solutions to issues around deplatforming or the denial of service to controversial websites.
• Reclassifying elements of the internet as common carriers would have a negative impact on innovation by increasing regulatory barriers to entry and limiting the choices services could make.
• Classifying social media and other elements of the internet as “common carriers” or otherwise requiring them to carry all content would also conflict with existing court rulings regarding internet speech.

Introduction

As the internet’s role increases in our lives, so too have debates over the laws governing it, including those related to online content moderation. Recently some scholars, including law professor Richard Epstein, have begun to argue that various elements of the internet ecosystem have become so indispensable and powerful that they are essentially utilities and should be designated as “common carriers.” They argue that such an approach is needed to address concerns about the ability of some companies to prevent access to certain websites or services.

This is not the first time such arguments have been made regarding various elements of internet infrastructure, nor is it the only portion of the internet that some have proposed to reclassify as common carriers or utilities. Some argue that social media platforms themselves should be considered common carriers and required to carry all users’ content, and others have argued that the Internet more generally has become a necessary utility requiring more forceful regulation. Given the existence of viable competitors such as Parler to the primary social media companies, however, the focus of this argument has shifted to companies such as Amazon Web Services that host websites and provide access to the internet.

In general, these arguments that elements of internet infrastructure are better understood as a utility or common carrier rest on a poor analogy given the current diversity and competition in the market. If such a change in the approach to internet governance were enacted, it would have significant consequences that could chill innovation and speech.

Why Are These Arguments Occurring?

In early 2021, the emerging social media app Parler found itself offline and unavailable to users after its removal by various app stores and the decision of Amazon Web Services to suspend its cloud service to Parler. Some scholars and commentators questioned whether this ability to effectively remove a service’s presence from the web is appropriate. They argued that the firms offering services that are part of the fundamental ability to have a voice heard online  should be viewed in a common carrier framework rather than as a private actor in competitive markets. Coupled with the decision of several social media platforms to indefinitely suspend then-President Trump’s accounts, some have argued that this growing power is beyond that typically permitted for private actors and is instead more akin to a utility. They allege that decisions that more directly affect the ability of websites to remain online, such as Amazon Web Services’s decision no longer to host Parler, are different than decisions by individual social media platforms, as an entirely new platform effectively have to build its own internet infrastructure in order to remain accessible if backend service providers refuse to host it. This significant power to keep an individual or service offline, they argue, should shift the governance of essential internet infrastructure toward a common carrier model and require these services to carry all content.

This description, however, does not accurately reflect the internet ecosystem or the current situation. There are many elements that are part of the online experience. These include the internet service providers (ISPs) that provide a user or service access to the internet, the cloud servers and webhosts, and other services that provide cybersecurity protection from hacks. As a whole, these services are typically referred to as the “internet stack.” For a particular user or service additional elements may be included as well, such as an online service such as a search engine, a social media platform, or an app store for mobile access.

For a common carrier approach to be appropriate, there would need to be a natural monopoly that would prevent the access to otherwise essential services. When it comes to the various elements needed to provide online services no such natural monopoly exists in the markets associated with any of these components. While there are distinctions about what a particular service provider does to keep a website available and its particular role in a website’s ability to stay online, the decision to deny service is often the result of a business decision or contractual violations. The reality is the resulting difficulty in finding alternative service providers may be due to similar terms, similar market forces, and the societal pressures all services face regarding a controversial or distasteful product. While there are large players at various levels of the stack, they are not the only option that a user or online service has to reach consumers. For example, Parler has since come back online by negotiating with another cloud host. The decision to deny service can be better understood as a contract violation and risk analysis rather than as evidence of a natural monopoly. While some argue that the power webhosts or other elements have has allowed them to function as gatekeepers, there are still competitive markets with choices and distinct terms of use from which a new service can use. As a result, the consequences stemming from termination of service may be due to poor business decisions or a poor match. The impact of these decisions does not necessarily indicate that a single player has become a monopoly, but does illustrate that many elements are involved in most websites.

Consequences of Common Carrier Regulation for Innovation and Speech

Common carrier regulation of social media, webhosts, or other portions of the internet stack would have several consequences that would be detrimental to consumers. It could deter innovation by locking in existing giants and increasing regulatory costs for both platforms requiring approval and the essential companies that the government would now have to micromanage. Such a change would also contradict the current application of the First Amendment to the internet and the rights of private actors.

Currently, platforms face market incentives to constantly improve and evolve lest they find themselves replaced by newer competitors. This is not the environment faced by a utility or monopolist. To classify various elements of the internet stack as common carriers would be an immensely costly regulation, as University of Florida Professor Mark Jamison has discussed regarding previous arguments about Google.   Such a classification would require intense government intervention into very granular business decisions of various elements of the internet such as the formation of contracts for service and decisions regarding which services can be bundled together. The result would not be to spur additional services through competition but to make it even harder for new players to provide an alternative or improved service by increasing regulatory barriers to the market. Additionally, existing platforms would be less likely to engage in improvements as they would have the guaranteed success that could be associated with having regulatory protection as a utility and a decreased likelihood of competition.

Beyond deterring innovation, designating webhosts or social media as common carriers is also misaligned with existing jurisprudence. As Santa Clara Law Professor Eric Goldman notes, many of the elements of the internet infrastructure under discussion have multiple competitors and courts have rejected the analogy that internet services are like broadcasters and should be subjected to equivalent scrutiny. Applying “must carry” requirements as part of reclassifying these services as common carriers would change this precedent in such a way that would end in much more government intervention into the speech decisions of private actors. Rather than being able to change services as a result of a desire for better terms or different standards, companies would have the content they must carry dictated by the government.

Conclusion

The various elements of internet infrastructure involved in allowing platforms to reach consumers reflect neither a natural monopoly nor aspects characteristic of a utility. Even if one disagrees with the decisions of private actors to deny service to those who violate their terms of service or contract requirements, classifying elements of the internet infrastructure as common carriers or utilities is likely to lock in existing giants and render new entry nearly impossible. Such a shift would set a concerning precedent for intense government intrusion into business decisions and change current interpretations of how the First Amendment interacts with these decisions. It is likely these debates around how best to regulate the internet will continue, but in considering any dramatic changes such as common carrier regulation, policymakers must understand the interaction between different elements of the internet and the evolving markets in which they occur.