Algorithmic Visibility and the Standardisation of Cultural Narratives
Over the past two years, the way people access information and cultural content has changed significantly. Artificial intelligence tools are now part of everyday use, acting increasingly as a first point of access to knowledge.
At the same time, many cultural narratives remain unevenly documented and difficult to access, particularly those that exist outside dominant linguistic and cultural contexts.
The Swiss Arab Cultural Alliance was launched within this evolving landscape, working on the circulation of contemporary Arab visual narratives that are often rich and diverse, yet still under-represented and unevenly accessible.
As digital and algorithmic systems shape how knowledge is accessed, they also influence which narratives become visible, how they are structured, and from whose perspective they are understood. This shift does not only affect access to information. It gradually reshapes how reality itself is perceived.
Visibility and its limits
Artificial intelligence systems rely on existing digital content, but this content is far from evenly distributed. Out of the approximately 7,000 languages spoken worldwide, only a small fraction are meaningfully represented in digital environments, and an even smaller number are supported by AI systems in ways that allow for nuance and depth. English and a limited number of dominant languages continue to structure most of the available data.
This imbalance is not only linguistic. It reflects broader dynamics that shape how knowledge is produced and circulated across contexts. It is also linked to the environments in which these technologies are designed. The development of artificial intelligence has largely taken place within specific geographic, economic, and social contexts, particularly in the United States and Western Europe, and within relatively homogeneous professional environments. Research has shown that the technology sector remains predominantly male and concentrated within a limited set of cultural and educational backgrounds, which inevitably influences how systems are conceived, trained, and deployed.
As a result, the way information is structured and made accessible reflects these underlying conditions.
From access to perception
As some narratives become more visible than others, they begin to shape how knowledge is perceived. Repeated exposure to similar perspectives influences how cultural contexts are understood, often in ways that remain implicit.
Certain narratives continue to exist but circulate less widely, or in altered forms. Translation, for instance, can make content more accessible while transforming its meaning, particularly when cultural references, nuances, or modes of expression do not fully carry over. In a similar way, visual languages can be detached from the contexts in which they were created, becoming recognisable styles without retaining the narrative depth that originally gave them meaning.
What is at stake is not only the presence of narratives, but the way they are interpreted. The conditions under which content circulates shape how it is understood, and ultimately which perspectives become more familiar than others.
A gradual standardisation of imaginaries
Within this dynamic, a more structural shift begins to emerge. As access to knowledge is increasingly mediated by systems trained on uneven and repetitive data, there is a tendency towards convergence in how narratives are presented and perceived.
Narratives do not disappear, but they become more uniform in form. Differences are often simplified, reformulated, or aligned with dominant formats that are easier to process and circulate across contexts. Over time, this affects not only what is visible, but the range of perspectives that are actively encountered.
This process does not eliminate cultural diversity, but it can limit how it is expressed and accessed in practice. When narratives circulate in more standardised forms, they no longer carry the same depth of context.
What begins to narrow is not only visibility, but the plurality of ways of thinking that narratives have historically made possible.
Beyond access: the role of narratives
Many current responses to these challenges focus on increasing access, whether by expanding datasets, supporting more languages, or improving inclusivity in technological systems. These efforts are important, but they do not fully address how meaning is shaped and transmitted.
Narratives have always played a central role in this process. Through stories, literature, and visual expression, societies have developed distinct ways of understanding the world, rooted in specific histories, references, and cultural contexts. These narratives convey information, and they carry ways of relating to the world.
This is precisely what is affected when narratives circulate in more standardised forms. As they become detached from their contexts, they may remain visible but no longer convey the perspectives from which they emerged. What is altered is not only their content, but their capacity to sustain different ways of thinking.
The issue is therefore not only access to content, but the conditions that allow narratives to retain their depth. Preserving this requires attention to how stories are documented, contextualised, and connected, and to who is involved in this process, so that they continue to carry the perspectives and intentions from which they emerge.
A concrete tension for cultural actors
For artists, publishers, and cultural organisations, these transformations are already part of everyday practice.
Visibility has become a necessary condition for circulation. Without it, work struggles to reach audiences, build recognition, or create opportunities. At the same time, making work visible also means exposing it to forms of reuse that are difficult to control. Content can be integrated into datasets, reinterpreted outside of its original context, or circulated in ways that detach it from its authorship and meaning.
This dynamic is now at the centre of ongoing debates around copyright and intellectual property. As creative works are increasingly used to train artificial intelligence systems, questions arise regarding consent, attribution, and compensation. For many cultural actors, this is not only a technical issue but a structural one, as it directly affects how their work is protected, and valued.
In this context, the challenge is not simply to remain visible, but to ensure that visibility does not come at the cost of meaning. Cultural actors are increasingly required to navigate this balance, finding ways to share work while preserving the perspectives and contexts that give it depth.
The new Swiss AI Apertus focuses on openness and transparency. © EPFL/ETH Zurich/CSCS/molinari design
A Swiss context
In Switzerland, initiatives such as Apertus, developed by EPFL, ETH Zurich and the Swiss National Supercomputing Centre, reflect a growing effort to address some of these challenges. By focusing on multilingualism, transparency, and open-source development, these projects aim to broaden access and reduce dependence on dominant technological infrastructures.
These developments are essential, particularly at the level of infrastructure. They expand what can be processed and made accessible, and contribute to a more diversified technological landscape. However, they do not, in themselves, determine how narratives are contextualised.
This approach reflects a broader effort to rethink how AI systems are developed and governed, as highlighted by the Apertus project:
“Apertus is designed for the common good. It is one of the few large language models of this scale to be fully open source, and the first to integrate, from the outset, core principles such as multilingualism, transparency, and regulatory compliance.
”
Connecting existing efforts
Across the Arab region, similar questions are being addressed through academic and institutional collaborations. Universities and research centres are increasingly working on artificial intelligence, language technologies, and digital humanities, with the aim of better reflecting Arabic language and cultural contexts within technological systems.
Some of these collaborations are already taking shape. In 2025, the American University of Sharjah and Al Akhawayn University in Morocco formalised a partnership bringing together Arabic studies, digital humanities, and the ethical development of emerging technologies, including the creation of an open-source Arabic dataset.
At the same time, investment in artificial intelligence and data science is growing across the region, pointing to the emergence of a dynamic ecosystem engaging with questions of language, data, and representation.
At this stage, a broader question emerges: what role can literary, editorial, and narrative practices play in ensuring that stories remain situated, and continue to carry the perspectives from which they emerge?
The role of Swiss Arab Cultural Alliance
The Swiss Arab Cultural Alliance operates within this space. Its approach focuses on engaging with stories at an early stage, before they are reduced to data or detached from their environments. By working with artists, publishers, and researchers, it seeks to ensure that narratives retain the perspectives from which they emerge.
The Alliance operates at the level where narratives are shaped, not only where they circulate, contributing to forms of visibility that preserve their complexity while connecting cultural, academic, and technological ecosystems across Switzerland and the Arab region.
Reframing the question
As artificial intelligence becomes more widely embedded in everyday systems, questions of ethics and representation are becoming increasingly central.
For under-represented voices and narratives, this is not a secondary issue. It requires engaging with artificial intelligence as an integral part of the process, rather than treating it separately, as doing so would limit the real impact of how these perspectives are carried and understood.
The challenge, therefore, is not only to make these narratives visible, but to ensure they are meaningfully integrated into the algorithmic systems that increasingly shape access to information, while remaining grounded in the contexts and perspectives from which they emerge.
