Data Concealment and Lack of Transparency: Do Online Platforms Enable Research Into Their Operations?

Major online platforms are key actors in shaping the public sphere, yet many aspects of their operations remain insufficiently researched. Instead of openness and cooperation, the academic and research community is faced with complex procedures, high costs, and non-transparent policies and rules that prevent a deeper understanding of how platforms operate.

Photo: Zašto ne

For the purposes of this article, interviews were conducted with Marta Mitrović (Faculty of Philosophy, University of Niš), Feđa Kulenović (Faculty of Philosophy, University of Sarajevo), and Mila Bajić (SHARE Foundation).

In October 2025, Google removed years’ worth of data on political advertising in the European Union from its official websites. The data had been collected during numerous elections held in EU member states and were of great importance for conducting various studies on political campaigns and the activities of different actors.

This move further complicated the work of researchers, who lost access even to data that had previously been publicly available. All this is happening in a context in which major online platforms already make access to data difficult for those who wish to use it for research purposes. Platforms, largely in an effort to protect their own interests, remain significantly non-transparent in this aspect of their operations.

Conducting research into certain aspects of the work of major online platforms may reflect negatively on the platforms themselves. It is clear that a lack of transparency benefits their business models: research on the protection of user data, for example, shows that users largely distrust platforms when it comes to the protection of their privacy, yet still accept platform terms, because otherwise they would have to stop using their services.

And while this issue could partly be attributed to insufficient personal responsibility on the part of users when it comes to sharing personal data, the greatest responsibility nevertheless lies with platforms whose terms of use are overly extensive and excessively complex to understand. What is even more problematic is the lack of insight into how and for what purposes users’ personal data are used. Scientific research makes it possible to highlight these and many other problems and shortcomings in platform operations, which ultimately runs counter to platform interests. This is precisely where the motivation for such an attitude toward the academic and research community may lie.

What does the work of researchers from our region look like in practice?

Major online platforms, in most cases, avoid enabling easy access to their data, thereby maintaining non-transparent practices. Some platforms have even discontinued tools that were previously publicly available. In August 2024, Meta shut down CrowdTangle, an analytical tool that enabled researchers, journalists, and others to monitor social media, collect information on disinformation and other potentially harmful content, free of charge and without registration. It was replaced by the Meta Content Library, access to which can only be requested by academic and non-profit researchers. However, access to data from major online platforms is often a very demanding and lengthy process for researchers. In addition, it can also be expensive.

As our interviewees shared with us, the process of verifying eligibility for access to data from major online platforms requires researchers to provide extensive documentation and takes a great deal of time. Furthermore, access to some data requires special, paid tools or software, for which domestic academic and other institutions are often unwilling or unable to allocate funds.

As a result, some researchers give up on studying this highly important field. Consequently, this leads to insufficient research into certain aspects of the operations of major online platforms, particularly in smaller regions and language areas. One such region is the Western Balkans.

Researchers from our region are in a less favourable “negotiating” position with online platforms, as platforms often have no particular interest in individual, relatively small countries. Experience has shown that major online platforms are more inclined to act and consistently implement their own rules and policies in countries that belong to larger language areas, such as those where English is spoken (1, 2, 3).

Therefore, data collection and analysis are crucial for proving irregularities and harmful effects of online platform operations in the Western Balkans, as well as for advocating improvements. All of this is, however, made more difficult by the non-transparency of the platforms themselves, as well as by the lack of recognition of the importance of researching this field within domestic institutions.

A key aspect of the value of research conducted on local samples and within local contexts is the improvement of content moderation mechanisms of major online platforms. Moderation of hate speech is particularly significant. As our interviewees pointed out, current mechanisms are unable to recognise local context and adequately moderate content. Algorithms only “see” what is written, while in many cases context and implied meaning are crucial for determining whether something constitutes hate speech.

As a result, numerous reports concerning content that people from regional countries who share the same language, as well as certain social and political contexts, know it is hate speech remain without adequate platform response. Responses either fail to arrive altogether, or moderation mechanisms “determine” that no violation of policies has occurred. What is missing, therefore, are moderators; real people from the local environment, who could ensure that such reports receive appropriate treatment.

The functioning of algorithms, whether in the context of moderating harmful content or in the creation of “filter bubbles” or virtual spaces in which we are exposed only to views and opinions similar to our own, remains largely unknown. Without adequate research, evidence of their harmful effects remains merely anecdotal.

What is crucial for researchers from the Western Balkans?

Given that each Western Balkan country individually represents an insufficiently large market in the perception of major online platforms, it is crucial for researchers in the region to connect and conduct joint research on issues characteristic of the entire region. In this way, platforms can be approached in a unified manner, and the size of the collectively perceived market becomes more significant.

Researchers also need support and networking. In research on online platforms, an interdisciplinary approach and cooperation, for example between the humanities and technical sciences, are essential. Our interviewees also believe it is crucial for the academic community to act together with civil society and other actors. This would allow them to jointly apply for funding that is not accessible to academic institutions alone, while such research would also carry more weight, as data could be analysed more comprehensively. In addition, the findings of such research could reach a wider audience, thereby increasing the impact of advocacy for better conditions and regulation of major online platforms in our region.

New opportunities for domestic researchers to access platform data are also opened by the European regulatory framework for platform transparency contained in the Digital Services Act, which obliges platforms to enable access to data for research purposes under certain conditions. While this mechanism is mostly reserved for so-called vetted researchers, access to publicly available data is not limited to academic or research institutions and can also be requested by non-profit organisations.

What is particularly important for our region is that researchers requesting access to data do not have to be based in the European Union – the only requirement is that the data be used for research that contributes to identifying and understanding systemic risks in the EU. In the regional context, this connection is very clear: one example is elections in Croatia, where the spread of harmful narratives and disinformation originating from the region can significantly affect democratic processes.

Some of the mechanisms envisaged by the European framework are therefore already available to researchers from this region, particularly within joint projects with European scientific or academic institutions. By adopting legal solutions modelled on the Digital Services Act, this space could be further expanded.

What is already necessary now is to raise awareness of the importance of such research, the need for cooperation, and the available mechanisms for requesting data, whether those offered by the platforms themselves or those enabled by the legal framework. All of this could lead to increased interest within the academic community itself and greater openness of academic institutions toward research projects of this kind.

(Marija Ćosić and Maida Ćulahović, “Zašto ne”)