By Analice Paron* and Thais Lassali

*Analice Paron is a journalist and researcher. She holds a Master's in Social Anthropology from Unicamp and is a doctoral candidate in Sociology at UFSCar. Her research interests encompass pornography, market dynamics, discourse analysis, and social media platforms.

On the morning of October 20, 2025, with support from the Department of Scientific and Technological Policy at the Institute of Geosciences/Unicamp, GEICT hosted professors Thomas Poell and David Nieborg for the lecture “Artificial Intelligence, Platformization, and Cultural Production.”

Thomas Poell is Professor of Data, Culture & Institutions at the University of Amsterdam and co-author of “Platforms and Cultural Production” (2022) and “The Platform Society” (2018), both translated into multiple languages. He also co-edited “The Sage Handbook of Social Media” (2018), “Social Media Materialities and Protest” (2018), and “Global Cultures of Contestation” (2017). David Nieborg is Professor of Media Studies at the University of Toronto, having served as a visiting professor and fellow at MIT, the Chinese University of Hong Kong, the University of Amsterdam, and the Princeton Institute for Advanced Study. He is co-author of the books “Platforms and Cultural Production” (2021), translated into Italian and Chinese, and “Mainstreaming and Game Journalism” (2023).

The lecture, held in the auditorium of the Institute of Geosciences, was divided into two parts. In the first, Professor David Nieborg presented his perspectives on the relationship between cultural production and digital platforms. Nieborg’s argument stemmed from the impression, shared by users and theorists in the early years of the internet’s popularization, that digital spaces would help democratize access to culture and cultural production. However, as Nieborg argued, what occurred was an increased concentration of power and information, particularly with the emergence and growth of major platforms. Companies like Apple and Meta end up controlling, through their algorithms, the visibility regimes of the content hosted on them—a process that aligns the interests of these companies with the sociotechnical infrastructure they create and sustain.

Picking up the thread from Nieborg’s presentation, Professor Thomas Poell discussed his own research on major cultural producers within the global digital market. Poell argued that cultural producers must increasingly adapt their business models to those of major platforms to succeed in these markets. This alignment involves understanding the three elements that, in practice, constitute a platform’s operation: the market, the infrastructure, and the governance. Poell argued that platforms are, first and foremost, business models that create digital infrastructures, which become visible to the public (and the market) through the websites and applications we access daily. Governance acts as the intermediary within this ecosystem of platform-managed businesses. It constitutes the entire domain of decisions aimed at sustaining and expanding a digital system that is simultaneously embedded in our daily lives and fundamentally opaque.

The event brought together students interested in learning more about digital studies, platforms, and digital technologies. Later that same day, the professors joined GEICT members for a roundtable discussion. During this talk, they engaged with students on the group’s ongoing research and its various points of convergence with digital platforms and digital ecosystems.

The interview that follows was conducted collaboratively by journalists and researchers Thais Lassali and Analice Paron.

---

What comes to mind when you hear the term “platformization”? Perhaps you think of its usefulness in describing new labor models mediated by major technology companies such as iFood and Uber. Or you might employ it to analyze how social media—with their recommendation algorithms and terms of service—reshape the ways we build relationships. Alternatively, the concept may help clarify shifts in how we access and consume entertainment, including films and series. This multiplicity of meanings and applications can make the phenomenon appear diffuse, even confusing. In this context, it is more urgent than ever to clarify its possible definitions and, further, to examine its implications—not only for everyday life but also for research concerned with the digital.  For this reason, the GEICT blog took the opportunity of Professor Thomas Poell’s visit to Unicamp to conduct an interview exploring the various conceptualizations of platformization, the consequences of the link between platforms and the widespread adoption of AI, and broader perspectives on technology and culture. We hope you enjoy the reading!

Thais Lassali (TL): First of all, thank you for the great lecture you and David Nieborg gave this morning, and thanks as well for agreeing to talk with us. I’d like to begin with something that is not exactly “global” from the point of view of its origins. I want to ask about the current political crisis in the United States, since most global platforms come from there. What do you make of this global power shift, especially as the political landscape seems to be leaning toward fascism? Platforms play an important role in that process, so I’d like to know what you think.

Thomas Poell (TP): Sure. First, to clarify, when we talk about global perspectives, we’re not referring to a universal view from “the global,” but rather to perspectives grounded in specific places—São Paulo, Lagos, Tokyo, and so on. The idea is to disrupt the flattening of cultural and political relations and emphasize that people, citizens, and cultural producers are situated very differently. Global perspectives start from situated experiences, not from universal assumptions.

Now, regarding the rise of fascism in the US and elsewhere—Brazil has had its own episode, and Europe, including the Netherlands, is now also deeply affected by right-wing populism. This shift has become entangled with platforms in complicated ways. As I mentioned earlier, public communication has moved from tightly gatekept media—television, major newspapers—to platform environments that are also curated, but in different ways. This shift opensopportunities for right-wing populist expression, and these forms are often enabled by US-based companies. In that sense, platforms become vehicles for much wider populist movements.

This has been unfolding over the past 10–15 years. Initially, platforms were even seen as opportunities for progressive, left-wing movements. Over time, however, they were increasingly captured by the right in many places. What is especially concerning now is the alignment between a right-wing US government and platform companies, which have tied their own futures to that administration. As a result, many responsibilities these companies had begun to assume—around democratic governance, transparency, diversity—are now being abandoned. That creates deeply unsafe environments, not only on platforms, but also in the governance of AI models. You may have seen that OpenAI has relaxed some of its moderation policies—politically, that is dangerous.

This is why it is crucial for regional governments—especially those leaning toward democratic or progressive values—to push back and refuse to accept current platform moderation practices. Brazil is doing that. Europe is trying to do this, though it is under enormous pressure from the Trump administration, which opposes European intervention. This is a clear power struggle. If we care about democratic futures, resistance and pushback are essential. 

TL: Yes, totally. While listening to you today and reading your work, I wondered: do you think platformization could operate as a form of colonization? For example, we’ve seen US concerns about Brazil’s PIX payment system. Aren’t these [financial market] companies [that are pressuring the Brazilian government], in some way, trying to take away our sovereignty?

TP: Yes—absolutely. This is central to the work of Nick Couldry and Ulises Mejías on data colonialism, which argues that these companies continue colonial power structures, just in a different form. I think they make a strong case.

But I also want to emphasize possibilities for negotiation, resistance, and alternatives. Intellectually, it is crucial to recognize these power relations and historical continuities. Yet the field [of digital studies] must also think about how things can be different—how to change or resist platform power.

That includes regulating or blocking platforms in ways that align with national democratic rules and cultures, whether in Brazil, Europe, or elsewhere. It also includes stimulating local alternatives in the tech industry. Brazil has been doing this quite effectively. Europe has tried, with mixed results.

Analice Paron: What is the relevance of talking specifically about social media platforms when there are many types of platforms? What distinguishes social media platforms from others?

TP: Which other platforms are you thinking about?

AP: For example, Instagram compared with Uber or WhatsApp.

TP: Right. Social media platforms are especially interesting because they are huge—they have massive user bases—and they intervene directly in culture and public discourse. That’s why they are central to our research. But they are also very specific, and we can’t generalize from them.

For instance, social media communication is scalable: anything circulated on social media can reach global audiences. Uber transactions, by contrast, involve direct labor and are not scalable in that way. Social media platforms rely on advertising, whereas Uber operates through fees. Different business models produce different relations of dependency.

So, we need specific theories for specific platforms. Gig-labor platforms like Uber or delivery apps require a different analytical framework than social media or webcam platforms¹. Some platforms resemble social media—like webcam platforms, which are also scalable—but stigma and sexual content shape their market dynamics differently. They are governed differently as well.

In short: be attentive to the specific characteristics of the platforms you study. Situate them regionally. Platformization looks different in different sectors and places—it always takes particular forms.

AP: But the concept of platformization answers part of that question…

TP: Part of it, yes. But platformization, as we write about it, is a rather generic concept. It’s only a starting point. You always need situated research. The type of “market” a platform creates differs significantly: Uber, food delivery, sex work, social media creation—all are different markets with different relations, different forms of exploitation, and different forms of precarity.

AP: In your presentation, you and Nieborg divided platforms into market, infrastructure, and governance. Where do the Terms of Use belong?

TP: Governance. Definitely governance.

AP: Why not infrastructure?

TP: Good question. In practice, infrastructure and governance are deeply entangled. Infrastructure, as boundary resources², always contains standards and rules—forms of governance built into it. We separate them analytically so we can engage with different scholarly debates in governance studies and infrastructure studies. But in practice, they’re intertwined. The separation is an analytical artifact that helps us talk about these concepts more clearly. Does that make sense?

AP: Yes, of course. Another issue that seems entangled with these three dimensions is copyright. Can we talk about copyrights, royalties, and author rights in the context of generative AIs?

TP: That’s a huge topic. Here again, power and recognition will shape how authorship is handled. Initially, OpenAI and Google completely disregarded authorship—they ingested the entire internet, including copyrighted books, such as those in Google Books.

Recently, however, powerful rights-holders have pushed back. The New York Times has sued OpenAI; there’s another major lawsuit against Anthropic; and groups of authors have won cases leading to billions in compensation. But these are very powerful publishers. Their intellectual property is more likely to be protected because they can challenge tech companies in court.

For smaller creators, for example, a social media creator in Brazil whose content gets scraped, the chances of compensation are slim. It’s difficult for individuals to challenge these companies legally. Collective or state-level action is more plausible but still challenging. 

TL: And on the other hand—do you think artworks made with AI could be authorial?

TP: That’s a complicated issue. There is no direct copying happening, so it’s not a simple copyright violation. A person may craft a very original prompt that generates an original artwork. But to establish copyright, you would need to copyright not just the artwork but the style. Historically, that has been very difficult. And if it were allowed, it would have a chilling effect on creativity, because anything created “in the style of” would become restricted. That would block entire creative genres. So, I think that would be deeply problematic.

TL: Staying with creativity but thinking about aesthetics: generative AI depends on databases made from human artistic work. Still, GenAI seems unable to replicate the complexity of human aesthetics—it often produces something that “looks like AI.” Yet many people fear that AI might someday dominate the artistic landscape. Do you think that’s possible, or will AI simply become another kind of aesthetic?

TP: I’m not sure we need to think in such extreme scenarios. Part of our presentation aimed to move away from these extremes. I don’t think AI will completely take over entire creative industries, at least not soon.

What is more likely is partial integration into specific tasks—background coloring in anime, generating characters, voice simulation. These areas are vulnerable to automation. They threaten certain types of labor, such as background artists or voice actors. But they don’t replace high-profile artists whose work is more protected.

The transformation will be fragmented, hybrid, negotiated—not total replacement but messy coexistence.

TL: Great. That’s it?

AP: That’s it.

TP: All right.TL: Thank you so much for talking with us and for your thoughtful answers. I hope our blog readers enjoy this conversation as much as we did.

1. Webcam or camming platforms are digital environments dedicated to live video broadcasting, in which performers engage in interactive shows and are compensated primarily through tips, subscriptions, or pay-per-action models. These platforms constitute a specific segment of the erotic market, mediated by digital technologies. For further reading, see Lorena Rubia Pereira Caminhas and her work on erotic webcamming (2020), Cristiane Vilma de Melo and her research on the de-platformization of camming workers (2024), and Núbia Ramalho and her study on racial dynamics among camming workers in Brazil (2025). ↩︎
2. In this context, boundary resources refer to the tools and regulatory frameworks that act as interfaces between proprietary platform software and third-party application developers, facilitating content/application creation while allowing platform owners to maintain control over its structure and design. ↩︎