Human Rights Are Not a Bug

In June, a malfunction in the content delivery network (CDN) Fastly took down large parts of the Internet. The same happened weeks ago with an outage by another CDN, Akamai. These outages rekindled a discussion on the roles and responsibilities of Internet infrastructure providers that goes beyond questions about security and reliability, but gets into the human costs of corporate domination of our digital worlds.

As a researcher in Internet infrastructure, I recently published a comprehensive report in which I argue that human rights impact assessments should be mainstream in the design, operation, and maintenance of communication infrastructures.

Talking Back

I grew up in a small town in the Netherlands and was always drawn to my grandfather’s shortwave radio. By turning a dial I could hear everything - from people speaking in Russian, Arabic, Chinese, to opera and jazz - and all of a sudden not feel so isolated. Roughly fifteen more years would pass before I read a lecture by Bertolt Brecht where he said, “for the full potential of radio to be fulfilled, the listener should be able to talk back.”

I’ve spent my career working to extend this maxim to the Internet. My latest research centers on why an Internet user must have the ability to talk back meaningfully by shaping how the Internet works, and the implications for privacy, security, and access to information if we can’t.

The Internet is often described as a global network of networks, but if this network is truly global, why is it nearly impossible to have an email address in a non-Latin script like Chinese, Arabic, or Hindi? This is not just true for applications, but also for the protocols, programming languages, routers, and pretty much every part of the Internet’s foundation.

The Internet as we know it is built on an infrastructure shaped by these kinds of inherent biases, norms, and values, which in turn shape our information societies. The question is, who shapes this infrastructure?

The Shape of the Internet

Historically, global communication systems like the telegraph or the telephone were owned, operated, and governed by nation-states or intergovernmental bodies, such as the oldest still-functioning one, the International Telecommunications Union, established in 1865. This, however, is not the case with the Internet, which is governed through a myriad of multistakeholder bodies that are dominated by multinational companies that run, own, and provide the Internet’s infrastructure.

We should not forget that the history of the Internet started in October 1957 with the launch of the Sputnik satellite by the Soviet Union, amidst the cold war. The first human-launched artificial satellite emanated what Time magazine would call “those chilling beeps” that could be received all across the world.

This led to the establishment of the United States Department of Defense Advanced Research Projects Agency which would fund what later would become the Internet. Until the early nineties, the United States government would operate the backbone of the Internet, which actually prohibited the sending and receiving of traffic with commercial content because it was publicly funded through the National Science Foundation.

When the Soviet Union collapsed, the United States relinquished direct control over the Internet and started a process of privatisation and commercialisation, whereby power over the Internet transitioned from the US government to predominantly US corporations. Thus the Internet is the first global communication network where private corporations play a more important role than governments.

This power is exercised in the ownership and control over Internet infrastructure. Despite increasing amounts of consolidation, there is still a need for multinational corporations to collaborate, because none of them have, or should have, an absolute monopoly. These collaborations take shape through governance bodies, like the Internet Engineering Task Force (IETF), who develop the technical foundations of the Internet, such as designing an Internet that is resistant to surveillance by making encryption pervasive, and ensure that the Internet does not just function well on fast connections, but also on slower ones in areas where there is less connectivity. The prime directive of these actors is to increase the size of the Internet by enabling the connection of most networks and services possible. Currently, these bodies negotiate standards and rules that allow for 70,000 networks, browsers, operating systems, routers, networks, and applications from different vendors and providers to interoperate with each other.

Problematically, Internet governance organisations maintain a distinct governance philosophy: to be consensus-driven and resistant to centralised institutional authority over the Internet. Instead, they exercise a more informal power, setting voluntary norms for industry behaviour. But because Internet governance bodies are dominated by the transnational corporations who wield much greater power than others, and have more resources to produce technology and send people to participate in the governance process, these fundamental values leave the public interest dangerously neglected in governance processes. This means that norms and values that might decrease interconnection get resisted or subverted.

Concretely this means that the infrastructure on which we build our information societies is designed and shaped based on the interests of multinational corporations, not on the public interest. This needs to change.

Human Interests

With the increasing importance of the Internet, the technologies, organisations, and institutions that run it should be scrutinised based on their societal impact, their impact on human rights, on users around the world. We already see an increase in regulation and laws by governments in Russia, China, Brazil, Germany, France, the Netherlands, as well in the United States, exactly because companies are not taking responsibility to explicitly align their technologies and services with societal values, such as making it a choice for users through which organisations and countries data is routed and where it is stored. But these regulations will always lag behind need.

Only a few years ago no one would have expected the enormous role CDNs would have, whereas now they mediate most of the Internet traffic. This leaves them with the power to deny service to certain groups, while providing it to others. And with dire consequences if one of their services inadvertently fails, as happened not too long ago when many people could not access mainstream services, but also had problems scheduling Covid vaccination appointments.

Those who design, standardise, and maintain the infrastructures on which we run our information societies should assess their actions, processes, and technologies on societal impact, not just profits; the Internet, and the rights of its users, are simply too important. Crucially, governments and civil society should participate in these processes to hold these governance actors to account, including building in human rights impact assessments into their workflows, using the United Nations Guiding Principles as a template. Critically, we must reframe the Internet user as a citizen with rights, not as customers to be bought and sold. The rights and impact on citizens and society should be central in the design, standardisation, and operation of the Internet.

There has been a time when the Internet was mainly a project for the military, a time in which it was a project for academics, then came a time the Internet was a project for corporations. Now is the time we make the Internet a project for citizens and communities.

When a kid anywhere around the world tunes into the Internet, they should be able to feel what I felt when I sat in the glow of the tube radio: a sense of true connection. This should be a connection that speaks their language, is welcoming to their identity, an Internet that they can shape, and a light in which they can grow, experiment, and live. That is a human rights-respecting Internet infrastructure.