Regulating Social Networks Is the Tip of the Iceberg

The clash between Elon Musk and Brazil highlights a dilemma that goes much deeper than social media. At its core, it’s about the complexities of a sovereign nation regulating a digital space overwhelmingly ruled by US corporations. Preventing people from using their favorite social networks is obviously not a solution, but the broader transformations of the digital landscape in the past decades underscore that it’s not enough to just “let the market rule.”

In a 1993 interview for Wired, archconservative and tech evangelist George Gilder described the internet as exactly the “kind of metaphor for spontaneous order” envisioned by neoliberal economist Friedrich Hayek. According to Gilder, the web showed that “in order to have a very rich fabric of services you don’t need a regimented system of control.”

Three decades later, the reality of digital domination has put paid to Gilder’s fantasy about the spontaneous freedom of wired intelligence. Shaped by the profit motive, cyberspace has become an algorithmically controlled space where the rules of engagement are rigged by the interests of Big Tech firms. That space may technically be a “public sphere,” but it is one where freedom of speech is an illusion, since tech companies define which messages are promoted and which go unnoticed — or even blocked.

In that same sense, the conflict in Brazil is not just about regulating digital space or freedom of speech. Broadly, it concerns the possibility of any country to have a say in its digital destiny. Tax avoidance, the erosion of democratic deliberation, abuse of market power, and influence over crucial regulatory settings — such as corporate climate standards — are all sounding the alarm about the growing power of Big Tech. The good news is those concerns are being addressed in a series of antitrust cases on both sides of the Atlantic. The not-so-good news is that almost none of them get to the root of the existential threat that Big Tech poses for free societies.

The Not-So-Public Sphere of the Cloud

Law scholar Antoinette Rouvroy and her colleagues have coined the term algorithmic governmentality to describe corporations’ growing ability to shape individual behaviors and society at large. Worse still, the spread of algorithmic governance has only intensified our dependence on Big Tech services.

Amazon, Microsoft, and Google concentrate almost 70 percent of the so-called public cloud market. The term cloud is used to refer to computing services that you can consume but that are running on a distant computer. While some companies and institutions still keep part of their computing services in-house, migration to the public cloud is led — and controlled — by Big Tech. On these clouds, anyone can rent infrastructure, software, data and platforms as a service. In other words, Big Tech clouds are not simply spaces to buy a black-box technology service, as one would a piece of software. These clouds function as public marketplaces where thousands of other companies, from start-ups to tech consultancy firms, offer computing services.

As ever more organizations begin to operate from the cloud, the “Big Tech Three” gain what could be called a “panopticon” view of global capitalism. Amazon, Microsoft, and Google oversee the whole “computing stack” — the networked infrastructure and operations that make up the cloud — since every piece of digital technology runs on and relies on a space they control. This panopticon view is crucial for gathering information that gives them extraordinary market leverage.

In addition to high market concentration, powerful lock-in mechanisms also make interoperability — i.e., intersystem communication — almost impossible. The result is a bottleneck in a fundamental piece of global capitalist production. The world was reminded of this on July 19, when Microsoft suffered a massive outage after its partner CrowdStrike made an error during a routine software update: operations across industries, from major airlines to banks and health care systems, came to an abrupt halt.

Despite their rhetoric of innovation and competition, Big Tech companies are essentially intellectual monopolies. They capture value from the assetization of knowledge and data. In that sense, they are the ones getting richer as companies develop computing services on their clouds. In fact, cloud-using companies pay twice: once to use the cloud services needed to develop their own technologies, and a second time when they pay a fee to the cloud provider each time their services are purchased on that provider’s marketplace. Big Tech are also knowledge predators, sucking up the results of collaborative research (over 80 percent of their thousands of scientific publications) to enclose it for their exclusive profit (they share the ownership of less than 1 percent of their patents). The deployment of AI tools is only pushing the logic one step further, by automating the organization of social labor.

Sitting atop huge data sets harvested from their multiple platforms and services, these new digital giants benefit from powerful scale and scope economies. As a result, they can charge a substantial levy on the essential services they provide. The surge of their weight in stock market indices — almost 50 percent of the S&P 500 during the summer — reflects investors’ confidence in the ability of those firms to capture a growing share of future profits as a result of the generalized dependency on their products.

So long as governments do not challenge this process of monopolization, Big Tech can confidently project its ongoing economic and political dominance. However, there is a growing chorus calling on democratic governments to regulate these companies, make them pay a fair share of taxes, and begin to envision a digital alternative that puts the people and the planet first.

The Odds of Challenging Intellectual Monopolization

In late July, Brazil launched a state-run AI Plan. It consists of a set of action items aimed at expanding Brazil’s digital sovereignty and domestic digital ecosystem thanks to enhanced public and local computing power, cloud capacities, and a stronger regulatory framework.

Building a digital ecosystem that is not dependent on foreign companies is particularly urgent in Brazil, which has the largest community of start-ups in Latin America. Some of those start-ups have already fallen into the hands of Big Tech, such as Akwan, a powerful search engine developed by two professors of the Computing Department of the Federal University of Minas Gerais, which was acquired by Google in 2005. Others operate independently and enjoy promising prospects. WideLabs develops large language model (LLMs) that create autobiographies of Alzheimer’s patients by recording, interpreting, and organizing their memories. Another leading initiative is the Center for Artificial Intelligence (C4AI) at the University of São Paulo, which is developing LLMs for, among other things, the analysis of indigenous languages.

These initiatives highlight the skills and talent of Brazilian scientists and engineers. But they also show the difficulty of defying digital corporate power today. While ostensibly independent, WideLabs trains and hosts its LLM on Oracle Cloud; C4AI, housed in the public University of São Paulo, is a joint venture with IBM. Because developing AI requires a colossal amount of data and processing power to train and run models, the push for AI adoption leads governments and nonstate organizations to blindly accept Big Tech clouds. Yet developing an AI model or any other computing solution in the cloud means to be locked in. It becomes virtually impossible to dismantle all the software architecture.

The same happens to state operations as they migrate to the cloud, from everyday services like management of hospitals and schools, to entire national defense systems — everything will be locked in by Big Tech. During Jair Bolsonaro’s administration, for example, Brazil’s national statistics office faced pressure from Big Tech to conceal how and what type of citizens’ data was being collected. As  explained, Bolsonaro’s government welcomed tech’s push to introduce Big Data and AI into Latin America’s national statistical offices through public-private partnerships. Among those initiatives was the idea to replicate the e-commerce measurement methodology adopted by Statistics Netherlands (CBS) through an agreement with Google.

The attempt to monitor citizens’ e-commerce through state agencies ultimately met with opposition, but Big Tech does not admit defeat easily. In light of Brazil’s AI plan, Amazon accelerated its strategy to sign cloud agreements with public companies and public sector dependencies, tilting the scale against Brazil’s push for digital sovereignty and in open defiance of president Luiz Inácio Lula da Silva’s vocal opposition to Big Tech.

But beyond political rhetoric, it will be hard to scale back the power of Big Tech. The Federal Data Processing Service (SERPRO) is Brazil’s largest state-owned IT corporation, and it seeks to develop what is being described as a sovereign cloud. Yet, SERPRO has met with a group of Amazon Web Services (AWS) executives led by national security director Sean Roche, a former associate deputy director for digital innovation at the Central Intelligence Agency.

According to AWS, Roche’s job is to educate public sector officials about cybersecurity threats and cyber-readiness. Of course, that also means Roche will probably recommend the government to contract AWS services — which is, in fact, what he tried to do in his meeting with SERPRO. The story in Brazil recalls the fate of the European Union’s GAIA-X, a project that was intended to build an alternative to US Big Tech, a sort of European data ecosystem, but ended up asking US cloud giants for help.

Today the odds of challenging intellectual tech monopolies seem bleak. But we need to take those odds, daunting though they may seem. As a first step, we need to push for strict regulatory controls and for enhanced public access of data deemed to be of general social interest, while forbidding the harvesting of sensitive personal data. However, beyond the cases of United States and China, which are in a privileged position to regulate their digital ecosystems, a more radical alternative should be envisioned by other governments if they want to reassert national and popular sovereignty.

Considering the staggering advances of front-running firms, a truly public cloud solution could only be built internationally. The United Nations would be a natural place to gather the scientific and technical competencies needed to launch such a cooperative project, promising to provide universal open access to digital services on a global scale. Notwithstanding the current weaknesses and divisions within the UN, or the power of the Big Tech lobby, we should press forward and revive the old idea of computing as a public utility. Such international public utilities would require more than infrastructure, including other services like AI models.

With the exception of the United States and China, most countries have almost no say in their digital future. Yet there is plenty of institutional space and momentum to form an international coalition in support of a global cloud that works as a public utility, oriented to support the world’s development needs and respond to the planetary challenges of our times.