Publications:
Siri, Give Me Back my Eye: From Audio Culture to Video and Back (2019) DOI: 10.14236/ewic/RESOUND19.21
The Totalitarian Data Filtering of AI
This research topic examines the political, cultural, and epistemological consequences of artificial intelligence as a system of information selection, filtration, and synthesis. It focuses on what may be termed information parasites, disruptive or embedded informational forces that infiltrate media and computational systems, shaping how data is sorted, prioritized, suppressed, and recirculated. Within the context of AI, these parasites do not necessarily refer to malicious code alone, but to broader structures of bias, ideological capture, recursive contamination, algorithmic narrowing, and systemic distortions that influence how machine systems process reality.
At the center of this research is the argument that AI systems may increasingly operate through a totalitarian logic of selection, not because they always impose explicit censorship, but because they tend toward reduction, optimization, and the privileging of dominant or statistically reinforced responses. In this sense, the issue is not only misinformation, but the narrowing of interpretive possibility itself. The system appears to offer choice while continuously steering the user toward pre-filtered, convergent, and often normalized forms of knowledge. This produces what may be understood as an illusion of informational freedom, where plurality is simulated, but genuine epistemic diversity is weakened.
The research investigates how this condition emerged historically through the development of media systems, from mass media and mechanical reproduction to networked digital media, algorithmic personalization, and generative AI. It situates contemporary AI within a longer genealogy of technological mediation in which communication systems increasingly shape perception, reality, and political subjectivity.
This research topic approaches AI not simply as a neutral tool, but as part of a broader history of media technologies that organize perception and influence collective understanding. It traces a trajectory through several media formations:
Mass media and mechanical reproduction
Beginning with the age of industrial media, the research explores how mass production and mechanical reproduction transformed the image, politics, and public consciousness. Drawing on Walter Benjamin, it examines the shift from ritual value to political value, especially in relation to propaganda, reproducibility, and the mass distribution of visual forms.
Simulation and substitution in new media
With the rise of digital media, the research turns to theories of simulation and substitution, particularly through Baudrillard and Virilio. It investigates how reality becomes displaced by mediated models, and how virtuality can function as a space of implosion, where meaning collapses inward under the pressure of overproduction, speed, and spectacle.
Mobility, connectivity, and media augmentation
The research considers the smartphone and social media era as a decisive transformation in communication, where media becomes continuous, portable, and embedded in everyday life. In this context, events such as 9/11 may be understood as pivotal moments in the intensification of real-time mediated consciousness. Mobility and connectivity do not simply extend communication, they restructure perception through constant augmentation.
Fragmented media, big data, and AI
The current phase is marked by fragmentation, data extraction, algorithmic sorting, and generative computation. Here the research focuses on AI as a machine of synthesis that does not merely retrieve information, but reorganizes and re-presents it through probabilistic selection. This produces a new regime of epistemic authority, where the machine does not show multiple sources, as a search engine does, but increasingly offers a singularly formatted answer.
A central concern of this research is the distinction between search-based retrieval systems and generative AI systems.
Search engines generally index and rank external sources, allowing users to navigate multiple documents and competing perspectives, even if this process is itself shaped by ranking biases and commercial priorities. AI language models, by contrast, tend to synthesize information into a unified response. This creates a different structure of mediation. Rather than presenting a field of documents, AI often produces a linguistic surface of coherence, which can conceal omissions, exclusions, and ideological simplifications.
The research therefore asks:
How does AI filtering differ structurally from earlier search and ranking systems?
Under what conditions does synthetic coherence become epistemic reduction?
How do information parasites operate within training data, platform design, model optimization, and user interaction?
To what extent does AI create the illusion of neutrality while reproducing dominant norms?
What are the political consequences of replacing navigable informational plurality with generated singularity?
This research group engages a broad interdisciplinary framework drawn from media theory, philosophy, cultural theory, and critical communication studies. Key references may include:
Walter Benjamin, on mechanical reproduction, mass mediation, and the political transformation of the image
Herbert Marcuse, on one-dimensional society, technological rationality, and the illusion of freedom
Jean Baudrillard, on simulation, hyperreality, and the disappearance of the real
Paul Virilio, on speed, substitution, and the collapse of distance in mediated war and perception
Martin Heidegger, on technology as enframing, and the reduction of beings into standing reserve
Michel Serres, on parasites, noise, and interference within systems of communication
Lev Manovich, on cultural software, algorithmic mediation, and computational image cultures
Bruno Latour, on actor-network dynamics and the agency of human and non-human actors
Siegfried Zielinski, on media archaeology, variant technological histories, and non-linear media evolution
Together, these thinkers allow the research to analyze AI not only technically, but as a historical, political, and cultural apparatus.