In May 2024, the Saint Pierre International Security Center (SPCIS) launched the “Global Tech Policy at the Forefront” series, featuring conversations with leading experts on the impact of emerging technologies—such as AI, blockchain, biometrics, and robotics—on global governance and public policy.
On November 14th, we had the pleasure of interviewing Professor Cecilia Rikap. Professor Rikap is now the Head of Research and Associate Professor in Economics at the UCL Institute for Innovation and Public Purpose (IIPP). She is also a tenure researcher at CONICET, Argentina’s national research council, and an associate researcher at the COSTECH lab at Université de Technologie de Compiègne.
Professor Rikap’s research focuses on the global impact of intellectual monopolies, especially in the digital and pharmaceutical industries. She explores how these monopolies concentrate intangible assets, shape the distribution of profits, and fuel geopolitical tensions. She has written extensively on these topics, including two well-known books: Capitalism, Power and Innovation (winner of the EAEPE Joan Robinson Prize) and The Digital Innovation Race (co-authored with B.A.K. Lundvall). Her current work examines how tech giants are taking on state-like roles and their growing influence on global production, innovation, and knowledge systems.
Our interview with Professor Cecilia Rikap covers a broad range of topics related to the political economy of science and technology and the economics of innovation. It is divided into two parts. In this second part, she shares her deep insights on Big Tech’s dominance in AI and its profound implications for both national AI policy and global development.
Dr. Naikang Feng: In your recent work, you've analysed the strategies Big Tech companies use to organize their AI-driven corporate innovation systems. First, what makes Big Tech so different? How does it differ from other monopolized powers?
Prof. Cecilia Rikap: Among all the intellectual monopolies from different industries that I have studied, US Big Tech is clearly of a different kind. This is because of: (1) the sophisticated mechanisms they use to expand their corporate innovation system globally (with the exception of China and to some extent Russia, which have developed their own intellectual monopoly in this field, with Chinese Big Tech operating just like US ones but at a smaller scale); and (2) the type of knowledge that they have monopolized.
They control the whole AI global innovation network or AI stack, from research and coding to the development of apps powered by AI models. They do so by positioning themselves in parts of the whole AI stack that are prone to natural monopolization, creating bottlenecks and keeping a panopticon of the whole stack. And, unlike a patented drug that can only serve one—or very limited—purposes, AI can be used for the most diverse things. AI is not only a general-purpose technology but also an innovation method, one that is becoming more mainstream in the most diverse fields, from drug development to environmental sciences to social sciences. This means that a few Big Tech companies—in particular Amazon, Microsoft, Google, and secondarily Meta—have control over a new way of producing intangibles. This is of strategic importance for existing intellectual monopolies beyond Big Tech and is rendering other leading corporations a subject of Big Tech power.
Dr. Naikang Feng: In what ways do Big Techs control the AI field?
Prof. Cecilia Rikap: On the multiple sophisticated mechanisms they use to control the whole AI field, in my recent work, I show how they control the whole AI knowledge network, even sitting at the boards of the main AI conferences, how they use corporate venture capital to get privileged access to what start-ups are working on and steer that research according to their interests, and how they subordinate every organization in the AI stack through their clouds. Every model, sooner or later, will not only be trained but also run on their clouds.
The cloud is a KEY part of this story. And this is not just because of the material infrastructure (datacenters, undersea cables, etc.) that I describe in another paper with co-authors as a Means of Information and Knowledge Appropriation (MIKA). Of course, these MIKAs are indispensable because they are the material part of the intangibles that Amazon, Microsoft, and Google concentrate, but by themselves, they are not enough to explain these companies’ control of the global economy. Their clouds are also the universal platforms, marketplaces for purchasing the use of any computing service, not only AI. And these companies also have more data than anyone else and control the AI research networks.
Seen together, they control the three essential complementary products for developing AI: code, data, and computing power. If one would be concentrated or controlled without ownership by another actor, there would be a different—even if concentrated—distribution of power. But this is not the case. The same companies control the three crucial parts of the AI that we have today. This is not just a story of concentrating the “infrastructure”; it is a story about controlling the whole digital stack beyond ownership because they control the three mutually necessary pieces of it.
I suggest reading my open-access piece on this topic here: https://www.tandfonline.com/doi/full/10.1080/09692290.2024.2365757
and here:
Dr. Naikang Feng: In your recent article on the U.S. national security state and Big Tech, you describe their relationship as a "frenemy" one. How do different types of states, such as the US, China, the EU, and others, interact with Big Tech, and how does this relationship impact their ability to regulate these companies?
Prof. Cecilia Rikap: This question is very broad, so I will only reframe it. We cannot ask a question in terms of “states” in general without looking at the different types of states in the world. States are not homogeneous creatures. The US and the Chinese states are the most powerful states in the world. They are the ones that relate like frenemies to Big Tech because they need each other but they also compete for spheres of control, for setting the rules and norms that govern capitalism.
The story is different with the rest of the states of the world. Other core economies like the European Union could have larger space for regulating these companies and have little incentive to let them be, because US and Chinese Big Tech hamper the EU’s power in the global economy and do not bring any benefits (unlike the case of the US and Chinese state which benefit by having these giants as ambassadors of their respective economy’s global power).
So, the EU tries to regulate them but here precisely the US state is vocal in preventing this and pushes the EU against it. The EU also lacks an enforcement capacity to regulate and it is still to be seen if it can build an alternative that is not more of the same (more intellectual monopolies in the tech sector now coming from Europe).
If it is hard for Europe, imagine the rest of the world. The rest of the states are directly dwarfed by today’s rulers, thus Amazon, Microsoft, and Google. States become dependent on their technologies just like other organizations and the push for adopting AI has only made these cloud giants stronger. They are behind the hype, they push the accelerator pedal, and they are all in when it comes to selling that AI is a magical solution for whatever we need. In the process, we all become their subjects, even states.
Happy to share a paper that I wrote about this with whoever is interested because this one is not open access. Anyway, it is here:
Dr. Naikang Feng: In your case study of Mercado Libre, you mention "digital dependency" in Latin America. How does the concept of digital dependency update traditional dependency theory, and what role do regional platforms like Mercado Libre play in this dynamic?
Prof. Cecilia Rikap: Digital dependency refers to how, in that paper, we update the idea of dependency theory. It was originally developed to explain the existence of cores and peripheries in the world economy. Unlike theories of imperialism, that mostly make the capitals and states from the core responsible for the underdevelopment of the peripheries, dependency theory offers a more complex analysis. This asymmetry and subordination exist, but it is in part enabled by actors from the peripheries that are complicit in the value extraction conducted by core capitals from the peripheries.
What we see today is that regional platforms like Mercado Libre (or, for instance, Jumia) play that complicit role. Yet, there is a difference with the past, and this is that the complicit capitals of the past were laggards; they neither used nor developed frontier technologies. Mercado Libre, within certain AI niches and dependent on technology that it uses without access by purchasing cloud services from Amazon and Google, develops frontier technologies. Still, this technological catch-up did not contribute to developing Argentina, where the company comes from, or any other part of Latin America, where Mercado Libre operates. On the contrary, it has reinforced the value extracted from every actor participating in Mercado Libre’s platforms (e-commerce and fintech), expanding inequalities.
So Mercado Libre structurally depends on Amazon and Google and pays millions to these companies for using services on their clouds—such as services for querying and storing the data that Mercado Libre hoards from its platforms’ users—but only because this enables Mercado Libre to subordinate thousands of smaller players and profit at the expense of the value that these other players create.
More in the paper: https://onlinelibrary.wiley.com/doi/10.1111/dech.12839?af=R.
Dr. Naikang Feng: Finally, will the economic gap between developed and developing countries become larger in the age of AI? How can developing countries catch up with the productivity revolution enabled by AI?
Prof. Cecilia Rikap: The effects of intellectual monopolization go beyond controlling and profiting from specific discoveries. They reinforce core and periphery dynamics with the perverse twist of knowledge (and data) produced in peripheral regions being appropriated by corporations from core countries that afterwards sell resulting products, setting prices that include charging intellectual rents, also to those regions. This pushes the peripheries to keep on compensating for these unequal exchanges by deepening nature extractivism, further harnessing the world’s chances to solve ecological disasters.
At the industrial level, subordinate organizations compensate for the profits squeezed by intellectual monopolies by super-exploiting their workers and worsening working conditions, as in the known cases of sweatshops for big labels and fast-food restaurants’ franchisees. The super-exploitation of unrecognized workers of the so-called gig economy is a more recent example. The gig economy is populated by lower-tier platforms like Uber and Deliveroo that depend on Big Tech digital technologies and infrastructure. They compensate for the value lost in the hands of the latter by further exploiting drivers and delivery employees.
Ultimately, monopolizing how to learn and innovate could doom society to a perpetual split between a minority that knows and decides, and an ignorant majority structurally deprived from the capacity to learn and create. A better world will not come out of thin air. It needs to be planned, and that requires vision and imagination. Without the capacity to create and learn, all the latter becomes even less viable.
Comments