Algorithmic biases and the discoverability of digital cultural content

Abstract

The digital era has transformed how the production of culture is accessed, how it circulates, and how it is organized. In this article, I wish to discuss the notion of discoverability. This notion, I argue, is one of the most recent cultural policy instruments that has emerged in the digital era. Discoverability implies creating conditions under which the public can easily encounter (be proposed or offered) cultural content that is culturally relevant. In other words, the notion of discoverability includes the capacity to encounter local cultural content and content that is made in languages other than English. Discoverability, however, tends to function on algorithmic biases that privilege English-language cultural content and content produced by large global corporations. From a cultural policy perspective, discoverability is rooted in two basic dimensions: the regulation of culture and the accessibility of culture. This article emphasizes the place of French-language content and touches on two dimensions: the accessibility of French-language digital content in general, and the issue of cultural content from French-speaking minorities. In doing so, it also sheds light on strategies and policies that are pertinent to other languages and to other linguistic minorities.

Introduction

The digital era has transformed how culture is accessed, how it circulates, and how it is organized. Algorithms have created conditions where cultural consumption and cultural interactions are monitored and used to create knowledge about users, and strategies to privilege access to some content. These technical black-boxes also exclude or minoritize other content. The age of artificial intelligence has given rise to new forms of intermediation with arts and heritage. It is now possible to access a computer-generated lecture on topics as diverse as American minimalism in the visual arts and the history of French literature. Hordes of cultural content—often of poor quality—produced by artificial intelligence populate platforms such as YouTube with the purpose, one can only imagine, to monetize clicks and views. In its digital form, culture is not only produced and distributed through new techniques; it is subjected to those techniques. In the digital era, and through artificial intelligence, one may consider that culture is rendered to technique. Or to borrow from Martin Heidegger's thoughts, that it is enframed by technique (Heidegger, 1958), which makes of culture and cultural data a standing-reserve available to be further exploited. Vast digital archives and humanities material are now the raw material to concoct YouTube's top 10 best European cathedrals, novel of the twentieth century, most visited UNESCO heritage sites, and any other cultural item that can be listed in tens for edutainment purposes.

The digital era also comes with challenges for cultural policy as new technologies destabilize systems of production, but also systems for the protection of culture. Since 2020, reports on the implementation of UNESCO's 2005 Convention on the Protection and Promotion of the Diversity of Cultural Expressions include a special section, which states that parties to the convention can report their data on their population's use of streaming services. In 2024, France reported that 56% of its population subscribed to a digital cultural content provider such as Netflix, Spotify, or Amazon (France, 2024). Data presented by the United Kingdom indicates that 38.2% of the British population regularly streams music online, and that 13 million households subscribed to Amazon Prime Video, and 16 million to Netflix (United Kingdom, 2024). In Italy, the government reported that 72% of the population subscribed to one of many music and streaming service (Italy, 2024), a number that is close to what is reported by Canadian authorities, who declared that in 2023, 79% of the population had at least one subscription to a music or a video streaming service (Canada, 2024). It should be said that this number increased from the 70% that Canada had reported in its previous quadrennial report on the implementation of the convention.

This overview, albeit very cursory, of the use of streaming services to access cultural content in different countries speaks to the digital condition of the cultural world that is now increasingly consolidating. The fact that this dimension is an informational element that state members of the 2005 Convention are invited to monitor tends to confirm that even though this digitalization of cultural consumption is now more than 20 years old, it is still interpreted as a challenge to cultural policy and to national cultural industries. There are at least three major common grievances expressed by policymakers against global players in the music and video streaming industry. The first has to do with taxation and funding. For instance, in Canada, American streaming giants have been able to generate profits without being taxed, and without being required to reinvest some of these profits into national media funds that support local productions. Moreover, many of the revenues generated in advertising in conventional radio or television have now increasingly been diverted to online platforms, leaving aside fewer revenue-generation possibilities to local organizations. Beyond taxation, another area of contention has to do with moving audiences. The major streaming services have become increasingly popular across the globe, and they choose which content should be diffused on their own streaming services, and sometimes even produce the content. This has many implications for national cultural policies. Private and public national broadcasters have been either laggards or unequipped to deal with this situation, and their streaming services offers may not be as attractive as those provided by global streaming services. For national broadcasters, this situation means a decline in their audience, and for artists and producers, this situation creates risks as platforms may not be entirely interested in cultural productions geared to local audiences. This is a challenge to national cultural policies, which are typically designed around, and in support of these actors. Finally, the third grievance has to do with intermediation. New digital platforms may do some gatekeeping, and the hypothesis is that they may prefer content for a global audience rather than content that speaks to local or national realities and is made for a national public. But, more importantly, the technological infrastructure shapes how content is made available and how it is made known to the audience. Many fear that the algorithms and ways of ordering cultural content may be done in a manner that disadvantages some productions over others.

In relation to this last common grievance, expressed by different policymakers against global streaming services, I discuss the notion of discoverability. I argue that this notion, is one of the most recent cultural policy instruments that has emerged in the digital era. Discoverability implies creating conditions under which the public can easily encounter (be proposed or offered) cultural content that is culturally relevant. In other words, the notion of discoverability implies the capacity to encounter local cultural content, and content that is made in languages other than English. Discoverability, however, functions on the principles that algorithms are technical black-boxes that contain biases privileging English-language cultural content and content produced by large global corporations. From a cultural policy perspective, discoverability is rooted in two basic dimensions, and these are (1) the regulation of culture and (2) the accessibility of culture. As discussed here, this policy notion is the result of different social and political forces at play.

Trades and cultural sovereignty: Regulation of culture

The notion of discoverability is not a simple reaction to a changing technological landscape; it has deep roots in broader conversations about culture. In fact, the full history of the notion brings us back to the negotiation of the General Agreement on Tariffs and Trade (GATT) in the late 1980s and early 1990s. During what is known as the Uruguay Round (1986–1993), services had become a matter of negotiation, and this included cultural goods. From the perspective of the European Union and Canada, cultural goods needed to be considered separately as they pertained to cultural identity (Footer and Graber, 2000; Gagné, 2017; Sojcher, 2002). This perspective became an element of contention with the United States, where culture was often seen as a commodity, like any other, and should therefore not benefit from any special status. Operationally, what was defended in these rounds was the capacity of states to subsidize their cultural sectors and cultural industries. This dimension is absolutely essential to cultural policy. Should culture be considered a good like any other, this would mean that any subsidy to the cultural sector could be seen as an unfair advantage or as a breach in free-trade agreements, offering grounds for legal contestation and remedies through the courts. For instance, in Canada, postal service subsidies for Canadian magazines were targeted as being in direct contradiction with the basic principles of free trade (Armstrong, 2000). For other countries, subsidies to the cinema and audiovisual industry were in conflict with the trade environment that was emerging in the 1990s (Freedman, 2003; Parc, 2017; Schlesinger, 1997). The extent of limitations that emerged in addition to subsidies, tax incentives, licensing, or any regulatory restriction to a domestic market came to be considered an infringement on free trade and as an unfair advantage, and further, could lead to a court challenge. In other words, the evolution of trade agreements in the 1990s challenged the very possibility of cultural policy.

The response came in as an international coalition of States and governments that tried to carve out culture from trade agreements and tried to develop an instrument that would protect state capacities in cultural affairs. As reported by Graber (2006), the doctrine of cultural exception put forward by France, Canada, and the European Union was an interesting tool for public relations and created an impression in the media, but it did little in concrete terms to provide a framework for legal protection of culture in international trade. Between 1998 and 1999, the notion of cultural diversity emerged as a tool that moved toward elaborating a strategy to protect state capacity and sovereignty over the regulation of cultural affairs. As a multilateral organization focusing on education and culture, UNESCO was an ideal place to elaborate a resistance strategy. The 2005 Convention on the Protection and Promotion of Diversity of Cultural Expression is the institutional product of this long collective action that reunited many actors in support of the state's cultural sovereignty.

The long debate on the nature of relationships between culture, economy, and the state is where the notion of discoverability has emerged. There are two elements to keep in mind. The first has to do with time, with the duration of these debates in an era where technological innovations were elaborated as the place where culture in trade agreements was debated. These discussions happened at a time when the internet was increasingly democratized and making its way for domestic uses in the West. Additionally, new platforms on the web were created, enabling the broad sharing of music, videos, books, and other cultural material online. The forerunners of today's streaming services were making their appearance. And as the 2005 Convention was being discussed, the technological landscape of cultural industries also changed significantly. Unlike other UNESCO conventions, such as the 2003 Convention for the Safeguarding of Intangible Heritage, where the notion of intangible heritage was relatively clear to most actors, the scope of the 2005 Convention has expanded over time.

Interpretations, and the ways it is being implemented, extend how we understand it, and how the Convention is intended to be used as a normative and technical tool for culture. The changing technological landscape characteristic of the late 1990s and 2000s has created these unique conditions in which states have attempted to assert their sovereignty over cultural affairs and their capacity to regulate culture in a world of global free trade, while technologies emerged and disrupted how culture was being produced and how it circulated. The governments of Canada, France, and the province of Québec have played key parts in elaborating a sense of awareness around these changes. Discoverability is a result of these intergovernmental actions to assert cultural sovereignty. It should also be mentioned that the new digital content providers have also challenged numerous forms of taxation and cultural funding systems. In Canada, for instance, platforms such as Netflix were able to operate without the requirement to invest in federal media funds and operated largely in the margins of a funding system that also fuelled local content production. The purpose of cultural sovereignty is not only to retain the capacity to fund cultural production, but it also involves the capacity to structure a taxation system, and a funding system that can affect cultural enterprises and measures that concern the regulation of content.

The second element that in support of the notion of discoverability concerns widely shared social discourses in the 1990s about globalization. Globalization was seen as an opportunity, but it was also seen as a treat, and particularly a treat to culture. Many sociologists and anthropologists had already witnessed, then, how social discourses offered either a simplistic optimist view of globalisation as a world of economic prosperity for some, but also as a world where cultural differences would be leveled, and where minor cultures would rapidly disappear (Banerjee and Linstead, 2001; Matsuura, 2000; Tomlinson, 1999). Minority languages and local cultures would vanish and be replaced by a new global culture informed by market values and consumption. In advance of the new millennium, the social discourse of the 1990s was not immune to a dystopian view of culture, a view that had the allure of a sad evolutionist tale about societies, where some would grow while others would be condemned to disappear.

Whether or not these views were justified is hardly the point. It should be acknowledged instead that these social discourses had enough force to contribute to the problematization of culture that was convincing and preoccupying enough to ignite new strategies and new work towards the protection of culture. Problematization is meant here in the sense of a collective construction and negotiation of a problem, that is, with different roles or scripts that call for actions between different agents (Callon, 1980, 1986). The translation of this social discourse into a viable social problem to be discussed and approached by legitimate institutions, and its subsequent translation into a series of norms and principles guiding actions, was made possible through the networks of UNESCO. Between 1998 and 2003, actors connected to UNESCO brought the issue of cultural diversity to the global political agenda. In 1998, a conference was organized to promote the theme of cultural diversity, which led to a report (UNESCO, 1998) and to a declaration in 2001, which ultimately gave the necessary support to the Convention of 2005.

From an intellectual perspective, the notion of discoverability results from this trajectory—that is, from the social discourse on globalization that accompanied the reflections on culture and trade, to the articulation of these concerns into an intermediary notion such as cultural diversity. The same trajectory, in turn, influenced multilateral organizations and their guidance for cultural policy development.

On the one hand, discoverability is tied to the intellectual fabric of cultural diversity, and results from this reflection on the capacity to support culture against economic forces detrimental to minority culture, and the diversity of forms and practices of cultural expression. On the other hand, discoverability retains its nature as a socio-technical notion from a cultural landscape that evolved in a digital era. Can members of a given population find cultural productions that are both local and germane to their culture, and in their language? This question encapsulates the demands of champions of the notion of discoverability. The objective of discoverability is to develop tools to ensure that algorithms are ordered to make a user's language and local cultural content more visible in the proposition of music, movies, and books available on different online platforms.

Discoverability as accessibility

During the Fifth Meeting of Ministers of Culture of la Francophonie held in Québec City on May 24, 2025—, an event under the patronage of the Organisation internationale de la francophonie (OIF)—the ministers of culture produced a written joint declaration on discoverability (découvrabilité). The Declaration of the Fifth Conference of Ministers of Culture of la Francophonie acknowledges the important risks that artificial intelligence, algorithms, and technologies constituting the digital infrastructure pose for linguistic diversity and for the cultural creativity of Francophone nations and regions in the world (Organisation internationale de la francophonie, 2025). Citing the UNESCO Convention on the Protection and Promotion of Diversity of Cultural Expressions, and OIF's own digital plan, the declaration engages ministers of culture on six axes of collaboration, promoting the discoverability of French-language cultural content. The first has to do with the adaptation of cultural policies and their legislative framework to a digital environment. This includes measures such as requiring more transparency, interoperability, and non-discriminatory access to French-language content from online platform operators. Additionally, the intentions expressed in this declaration are not only to regulate, but also to offer assistance to artificial intelligence developers to educate them and facilitate the goal of discoverability of French language content. Second, the declaration also recognizes the importance of cultural cooperation between the global north and south to ensure that discoverability elevates every Francophone partner, including those whose contribution to digital culture is only emerging. This is an important dimension for fostering solidarity and true inclusion, in particular with regards to members from Africa.

African nations are well connected in the digital world, and many users access cultural content from their mobile phones. But how accessible is cultural content from African nations? OIF, one of the few multilateral organizations where African nations occupy a leadership role is obviously sensitive to the importance of cooperation. The third axis of the declaration is intended to reinforce capacity among both cultural policymakers, and also, the public. The intention is also to ensure that Francophone nations produce, through their education systems, enough skill to support the ambition of greater discoverability of Francophone content online. The fourth area of engagement of this declaration concerns support for the cultural industries in the digital era. The fifth area concerns the creation of new norms for digital content, and greater attention and care put on data in cultural projects. The sixth and final action item concerns public education, attention to French language cultural content, and greater public literacy on discoverability.

This declaration is the result of a long process of construction between different governments and member states of the OIF. In 2017, a first statement on the importance of discoverability had already been drafted in the Fourth Conference of Ministers of Culture of la Francophonie (Organisation internationale de la francophonie, 2017) in Abidjan. This declaration presented discoverability as a potential area of collaboration in cultural policy between members of OIF. The institutional Francophonie does not have a monopoly over discoverability; it is an issue that has concerned many different states and linguistic spheres, such as Portuguese-speaking states, which have expressed similar concerns over the years. The Comunidade dos Paises de Língua Portuguesa, an institution similar to OIF, has done important advocacy and cultural development work to ensure that Portuguese language and cultures have a presence and visibility on the web. However, the collective action that led to the creation of discoverability as a policy notion in multilateral organizations and in multilateral agreements, results from an impulse with roots in the Francophone World (Magis, 2023). Many francophone States supported the development of the 2005 Convention on diversity of cultural expressions, and years after their important efforts, the notion of discoverability emerged as a main area of cultural policy intervention for many French-speaking governments. In fact, in 2019, France and the Government of Québec developed a common strategy in support of greater discoverability of French-language content, and in support of a greater access to local content for their citizens.

Taking a closer look at discoverability and at these digital plans reveal two things, some of which are not entirely new. The first is the importance of the technical dimensions of the organization of culture that emerged, and its communication to the public. If it is true that the digital world of culture has brought us in spheres that were not typically common to cultural policy, and that adjustments are required to fully comprehend the nature and scope of changes, what is at stake is cultural intermediation. School curricula, arts marketers, shopkeepers, gallerists, journalists, art critics, and friends are only a select number of intermediaries that contribute to the knowledge of culture and that are now commonly encountered in our everyday cultural experiences. These intermediaries have been commonly seen as stakeholders in cultural policies. Art amateurs, and professionals of the cultural industries, as well as heritage professionals, have contributed to making culture available, and by doing so, offer an appreciation of their tastes or indications of artwork that can be seen as culturally significant (Jeanpierre et al., 2013; Jeanpierre and Roueff, 2014; Taylor, 2015).

Algorithms are human constructions, but they nonetheless intervene in the form of what Michel Callon would describe as actants (Akrich et al., 2006); they are sociotechnical intermediaries that recommend cultural work based on economic considerations, and predictive models. Discoverability aims to stabilize the identity and scope of actions of this new agent in the sphere of cultural policy. While algorithms and artificial intelligence are new, their role in the sphere of cultural policy is not. The second element that is revealed by discoverability is the question of what constitutes accessibility. The question remains central to cultural policy from a State perspective. Discoverability is, in a certain sense, the result of an intellectual heritage that brings us to some of the most important debates of the twentieth century regarding culture. That is the debate of accessibility, whose ancestry is encapsulated in strategies such as the democratization of culture, or even other strategies such as cultural democracy. Malraux's policy to democratize culture in France implied challenging the geographical and economic barriers to cultural content with the use of Maisons de la Culture (houses of culture) as a policy instrument to render culture more accessible to the general public (Malraux, 1966). Discoverability is an attempt to ensure that domestic content in languages other than English are accessible—online—to non-English speaking populations. Unsurprisingly, the digital age of cultural policy has shaped new discourse about cultural accessibility and participation (Casemajor et al., 2021).

Conclusion: Intermediation and intermediaries

According to Valtysson, tech giants and their platforms “[have the] power to not only form dominant discursive formations in marketising, but also in stating and communicating” (Valtysson, 2022: 796). They are well-equipped to circumvent discussions that are normally part of the democratic process. For this reason, cultural policy is in a difficult place despite a series of efforts put forward by many governments, whether it is to recuperate tax dollars, or to regulate content. That being said, concerted efforts—multilateral strategies, for instance—may prove beneficial and maintain state capacity over important areas of cultural policymaking.

Cultural sovereignty is an important component of state activities. While this debate has clearly sparked a strong interest in the Francophone World, the issues raised by digital platforms, algorithms, and artificial intelligence are far from limited to this linguistic and cultural group. In fact, new areas of cultural expressions open to new debates. For instance, in its latest report on the implementation of the 2005 UNESCO convention, the Government of Vietnam has reported that it was aiming to ensure that Vietnamese culture could be present and have a space in metaverses. Games and virtual worlds offer a new space of social interactions where language, cultural representation, and even representation of cultural heritage are issues of increasing importance for some.

From a theoretical perspective, I argue that the policy notion of discoverability opens some important, albeit common, areas of cultural governance. Discoverability is a socio-technical construct, a policy answer to a new form of technical agency that emerges, and has imposed itself as a powerful intermediary in making culture available. As of now, in the global competition for artificial intelligence and digital markets, Silicon Valley represents the space where these intermediaries are being formed, offering not only new cultural experiences and new ways of accessing cultural content, but also new ways of organizing culture. Governments are now interested in opening the technical black-boxes to negotiate new practices that will not further minoritize cultures on the web, and in platforms. Such attention is needed to dramatically emphasize the existing digital imbalance between languages and cultures in the digital world.