Brussels Effect Under Attack: Fatigue, Digital Fallacies, and Reinvention

James Görgen is Coordinator for Digital Markets at the Ministry of Development, Industry, Trade, and Services and Specialist in Public Policies and Government Management in Brazil.

In recent years, the world's most prominent technology companies have entered a kind of regulatory and judicial purgatory. Big tech firms are facing lawsuits on various topics in different jurisdictions. They are adapting their business models to comply with newly approved legislation in some countries. Simultaneously, they are trying to prevent this effect from spreading globally.

The European Union has launched the strongest challenge to the dominance of these conglomerates so far, approving several laws and establishing institutions to manage topics related to its digital agenda. Since 2010, a complex set of regional and national laws, regulations, and resolutions has emerged. The goal was to curb the advance of foreign conglomerates in digital markets and avoid negative impacts of their uncontrolled actions on European businesses and citizens.

The initiative has accelerated in the last six years with a regulatory package that includes, among others, the Digital Services Act (DSA), the Digital Markets Act (DMA), the Artificial Intelligence Act (AI Act), the Data Act and Data Governance Act, and the pioneering General Data Protection Regulation (GDPR), approved in 2016.

The Brussels Effect

Despite the high cost of its formulation and implementation, this intricate regulatory framework has come to be seen as a formula for export, just as European regulations in other sectors have set a global example. As a result, European Commission technocrats have been traveling the world meeting with policymakers and legislators interested in replicating these instruments with the same objectives. The initiative has been so successful that it has even received a striking name: the Brussels effect.

Brazil has not been immune to this phenomenon. Since 2018, several proposals inspired by or simply replicating the content of European texts have emerged. This is the case, for example, with Brazil’s General Data Protection Law (LGPD), which is based on the GDPR, and the recent draft bills on artificial intelligence, regulation of digital markets, combating disinformation, protecting children and adolescents in the online environment, and other lesser-known bills presented since 2020. Their texts, more or less explicitly, encompass European proposals, such as draft bills 2630/20 (Fake News), 2768/22 (Regulation of Digital Markets), and 2338/23 (Artificial Intelligence).

Articulated Reaction

A little over half a decade after the main package of European legal frameworks was set forth, the accumulated experience with this process has generated global attempts by tech firms and other opponents of regulation to prevent the Brussels Effect and the continuous export of policy to non-European countries. As researchers such as Anu Bradford, Mariana Mazzucatto, Tim O'Reilly, and Ilan Strauss show, attacks on the European model are anchored in two fallacies. The first addresses an alleged dilemma between innovating and regulating. Heavy state action and an excess of non-principled laws can delay technological innovation. Consequently, this could slow down a country's economic development. Complementing this argument, there are those who defend the business model of big tech firms, asserting that competition in the digital environment not only exists but is just a click away.

To better understand why these narratives can be considered fallacies, it is recommended to read the articles by Bradford and the trio from University College London (UCL). Bradford analyzes precisely the European Union process that led to the Brussels Effect and compares it with the near total absence of regulation in the United States. The three UCL researchers dissected Amazon's model to demonstrate that both users and companies operating within one of the world's largest e-commerce platforms find no incentives to seek other services. These parties end up being held hostage by a model built to privilege Amazon itself.

These two arguments are at the center of the criticism of the Brussels Effect by Christopher Yoo, a law professor at the University of Pennsylvania and founder of the Center for Technology, Innovation, and Competition, which has among its main funders Amazon, Meta, Google, Comcast, AT&T, and GSMA. On a recent visit to Brazil, as part of his global tour of several key countries in the debate on the digital agenda, which includes India, the researcher primarily attacked the DMA and the AI Act, arguing that both, if transplanted without criteria to any jurisdiction, can inhibit innovation and reduce competition.

Obviously, his counterpoint is the US digital market, which, among other reasons, grew and prospered without regulatory constraints in the digital field from the creation of its five global national champions, strategically developed as commercial and national security projects with significant public investment in the last 25 years. Yoo does not say, however, how small companies from countries that did not create their own big tech firms could compete on an equal footing with the giants without relying on some regulatory barrier that protects the expansion of domestic entrants in these markets. He also does not suggest any innovation promotion mechanism for companies that do not have research and development budgets like theirs.

Just to take an example, a study released by the European Commission earlier this year showed that if the EU is compared directly with the largest tech companies, it would lag significantly behind. For example, the top 5 US ICT companies spent the equivalent of €209.058 billion in 2022, a 28.4% increase from the previous year and a 688% increase from 2012. At the current pace, this will soon be double the total gross expenditure on research and development (GBARD) of all 27 European nations combined (totalling €117.368 billion in 2022). The Horizon Europe program has a budget of €95.5 billion for the 2021-2027 period, which is less than half the budget of the Big 5 for 2022. Amazon has a larger R&D budget than any European member state. The largest national government budget is that of Germany, the only EU country that ranks among the top five when states and private entities are compared directly. Alphabet, the second largest in R&D spending, has a budget more than twice that of France, the country with the second largest budget.

Despite the interested criticism and articulated attacks on the Brussels Effect, the regulation of digital platforms and AI companies is on the agenda of competition authorities and judges around the world. Recent decisions show how the lack of regulation throughout the 2010s, which resulted in the hypertrophy of digital "leviathans," has been reversed. In recent years, Google, Apple, Meta, Amazon, and Microsoft have led the most high-profile cases of abuse of dominant position, involving monopolistic practices, asymmetric relationships between app developers and the systems where they transact their products, unauthorized use of personal data for training AI models, as well as manipulation of algorithms to privilege the offer of products controlled by the system and investments in the capital of companies to maintain dominance in the market. In some episodes, ironically, the existence of laws anticipating ex-ante regulation is contributing to the companies themselves adopting anticipatory measures to avoid sanctions.

Global Regulatory Alliances

The global actions of these companies also require reactions at the same level and scope. This is causing unprecedented regulatory alliances to emerge. On July 23, the UK's Competition and Markets Authority (CMA), the European Commission, the US Department of Justice (DoJ), and the US Federal Trade Commission (FTC) published a joint statement on competition in artificial intelligence. The note focused specifically on the risks to competition and consumers presented by foundational models and AI products. The authorities summarized the threats to digital competition from the new technology in three main risks: centralized control of essential resources; increased or consolidated dominant power in AI-related areas; and financial negotiations between key players.

The alert from the competition authorities of both countries and the European Union requires a look beyond the development of the models, including the effects in the field of generative AI.

We are aware of other risks that may arise when AI is deployed in markets. This includes, for example, the risk that algorithms may allow competitors to share competitively sensitive information, fix prices, or collude on other terms or business strategies, violating our competition laws; or the risk that algorithms may allow companies to harm competition through discrimination or unfair price exclusion. We will be attentive to these and other risks that may arise as AI technology develops. In light of these risks, we are committed to monitoring and addressing any specific risks that may arise in relation to other developments and applications of AI, in addition to generative AI.

Two years before this initiative, regulators from some countries in Europe, Asia, Africa, and Oceania were moving around a network with objectives in another area: the protection of rights. The Global Online Safety Regulators Network brings together independent bodies to cooperate between jurisdictions. The work is done through the sharing of information, best practices, specialized knowledge, and experience, to support coherent and coordinated approaches to online security issues. The Network recently created two working groups to deepen understanding and collaborating on the topics of (1) Education and Awareness and (2) Technology.

The way countries regulate the digital ecosystem will affect the lives of billions of people. Regulating in an ethical and rights-based manner has emerged as a crucial challenge of our time. International cooperation is necessary to shape the normative forces that guide the digital ecosystem: governance, ethics, and regulation. There is no transnational group with a global mandate and a specific focus on regulating online security. The Network aims to fill this gap.

In addition to a global focus, the need to monitor the risks of digital laws requires a great effort to create regulatory systems and ensure the strict enforcement of detailed rules. A study conducted in June by the International Association of Privacy Professionals (IAPP) revealed the complexity of the tangle of legal intersections generated by the digital agenda regulations implemented by the European Union. The graph below shows that only the AI Act has points of contact with various other legislation and regulations that pass through areas such as security, labor, transparency, markets, accessibility, and data and consumer protection. The complexity of European AI law, approved this year, also required the creation of the AI Office, a unit with 98 employees and a budget of almost 50 million euros. It may seem like a lot, but the value is half the budget of its British counterpart.

Source: IAPP.org

European Digital Sovereignty

In addition to global alliances between regulatory bodies, another interesting alternative emerging in this scenario, once again coming from Europe, is to invest in programs and projects that guarantee the independence of big techs, going beyond mere protection of rights and defense of domestic markets through legislation. On September 24, a group of experts will meet at the European Parliament to promote the “Euro Stack.” The event, which gives its name to the project, will discuss the need to build structural solutions as a way to ensure sovereignty through investment by states in national digital economy ecosystems.

One of its initiators, Cristina Cafarra, a professor at UCL and co-founder of the Competition Research Policy Network of the CEPR (Centre for Economic Policy Research), puts the initiative in these terms:

How can we create an alternative to Europe's grotesque dependence on Big Tech infrastructure in the next EC mandate? Regulation is not the only answer. The 'digital' is a big part of the infrastructure that needs to attract large investments as an element of a much-needed industrial policy push, with democratic values. There are many great people working on this in Europe and around the world, it's time to move forward.

In an article in Tech Policy Press published last month, Caffarra went further. According to her, the European Union has led the regulatory process to control the advancement of technology companies in recent years, but has been losing ground to initiatives in practice in the United Kingdom and even in the United States. Caffarra recognizes that the Digital Markets Act (DMA) has been a significant step in curbing the abuses of economic power of the main global technology companies, but that there is almost a cult behind this normative. Due to its focus on company behavior rather than structural interventions, this legal framework would be failing to address the underlying issues of market concentration sufficiently.

For her, the need for structural solutions should consider more radical interventions, such as the breakup of large technology companies. This approach could facilitate a better application of regulations and effectively promote competition. In this sense, the practitioner recommends that the countries should take advantage of the current opportunity to collaborate with the US in combating the dominance of large technology companies, suggesting that bold actions are necessary to avoid negative consequences in the future.

Despite its early start, Europe has failed to advance the application of antitrust law in digital markets - the selection of cases has sometimes been eccentric, driven by complainants; cases have often been very narrow, focused on partial conduct and not on the overall "ecosystem"; damage theories have sometimes been inadequate, putting everything in the box of "exclusion" and not engaging with exploitation; remedies lack courage and foresight.

Next, we (as a European, I have the right to use the pronoun) adopted an ex-ante regulation – the Digital Markets Act (DMA) and its companion, the Digital Services Act (DSA) - in a frenzy of enthusiasm and self-congratulation (more Brussels Effect!), and in the hope that defining the rules of the game would induce large technology companies to behave well and comply with our European values spontaneously and good-naturedly. But the reality of implementing the law has already been revealed, and it is clear that, despite the initial optimistic proclamations, the law is far from self-executing. On the contrary, it is as difficult to apply as antitrust had previously proven to be.

This attempt to alter the status quo through normative instruments and regulatory actions is insufficient because it does not address all the points of the ecosystem controlled by these companies. As Maria Farrell and Robin Berjon have shown, the major US technology companies fundamentally control all the essential infrastructure for operating a digital society in Europe or any other region of the globe: content discovery, communications, identity, commerce, payments, data, and advertising.

Caffarra and her colleagues see in the so-called digital public infrastructures (DPIs), as India and Singapore have been betting on, and in policies to promote the local software and hardware industry, a way to break this logic.

Thinking About the Forest

The UCL researchers' charter of principles is ambitious. Would it be a way to try to reinvent the Brussels Effect now through the lens of industrial policy? Only time will tell. The fact is that in this field many countries have something to add – including Brazil, which has been taking great strides in the same direction in search of its digital sovereignty. The principle is present both in the Brazilian Artificial Intelligence Plan (PBIA) and in the Digital Transformation mission of the New Brazil Industry (NIB) or in the projects of Digital Public Infrastructure (Gov.br) and Sovereign Cloud that are being designed by the Federal Government.

But what we are doing differently in Brazil relative to Europe is not putting all our eggs in the same basket. Neither a tangle of complex regulations alone is the way out, nor is simply investing in a public cloud without other elements of an industrial policy. Neither just an AI plan will fill all the gaps, nor more competition alone will be able to oxygenate the digital markets and guarantee the entry of new agents. The digital agenda is much broader and needs to be very well sewn to truly put a country in a position to make some difference in establishing its national ecosystem with some kind of regional influence.

We need to understand that the issue of monopoly may be a non-issue for those private companies that may want to be dependent. The very division of competition versus monopoly does not make sense in current capitalism, organized as spheres of control that, in turn, are all interconnected, as shown by the Argentine researcher Cecilia Rikap. More than building a computational infrastructure under a national domain, it is necessary to directly attack the causes of intellectual monopolization, including the concentration of the so-called means of appropriation of information and knowledge.

The insufficiency of opting to attack a simple market instead of making a broader choice becomes very clear in the article, in which she analyzes how the artificial intelligence economy is already concentrated and monopolized by a handful of firms.

In the value chain of AI, the corporate armament of interdependence is supported by the combined concentration of intangibles, but also of tangible assets that provide the indispensable material base of intangibles, such as data centers. AI requires talent, processing capacity, and data.

Amazon, Microsoft, and Google, the cloud giants, are the only companies in the world with a high concentration of these three strictly complementary assets, which positions them in exclusive choke points of the AI stack or value chain. (...) I have shown that Microsoft and Google can use interdependence. as a weapon in AI research. This is just one of the networks where the centrality of Big Tech offers space for corporate network power, exercising forms of control that go beyond ownership.

Taking this broader scenario into account, a dialogue between the Global North and South to move in the same direction would be a way not only to strengthen the construction of common alternatives that exchange transnational oligopolies for domestic ones. It could constitute a legitimate effort to defend technological self-determination in a democratic way and without the need for isolationism or excessive bets on regulatory creativity. Instead of a Brussels Effect, why not invest in the impact of a South-North Alliance that will advance the co-creation of a better future, rather than the export of one region’s idea of what that future should be?

A Portuguese version of this post originally appeared here.