The origin of the term “platform” can be traced back to the Middle French “plate-forme”, which was defined as a level surface or a raised area along a railway track that enables passengers to access the trains. This definition emphasises the supportive role of platforms, serving as a structure for the operation of products or services, rather than creating something themselves (Belli, 2021). Likewise, the digital platforms we are dealing with in this article are auxiliary tools that do not produce anything themselves but provide an infrastructure for service providers and users to meet. However, unlike non-digital infrastructures such as train stations, digital platforms have potentially unlimited scaling concerning, for instance, their user base, data volume, technical features and geographic reach, rendering them more useful the more people use them (see on the difference between digital and non-digital platforms (de Reuver et al., 2018, pp. 126–127; Yoo et al., 2012).

For certain functions of social life, digital platforms have become the central places of exchange. Examples of such platforms include marketplaces (e.g., amazon, Alibaba, or eBay), social networking sites (e.g. Twitter, Instagram, or Tiktok), search engines (e.g., YouTube, Google, or Baidu), or payment systems (e.g., PayPal, Amazon Pay, or Skrill). In academia, we can also observe that research and its communication become more digital and that digital services are aiming to become platforms (Fecher et al., 2021; Owen, 2006; Pleskach et al., 2020; for an overview of the literature see da Silva Neto & Chiarini, 2023). However, what effects this looming platformisation brings in science and what potential benefits and harms follow, remain unclear.

Given the increasing significance of digital platforms in scientific inquiry and their influence on the social structure of research, it is imperative to formulate a robust conceptual framework of digital science platforms. The extant literature on scientific digital platforms, however, exhibits a narrow focus on particular scientific practices and lacks a clear-cut definition of this phenomenon (for an overview see da Silva Neto & Chiarini, 2023).

Here, we explore the concept of digital platforms and their potential impact on academic research. To clarify the concept of platforms, we first address the question: To what extent can digital platforms be understood as a specific type of research infrastructure? To this end, we draw from recent literature on platforms and platformisation from different streams of scholarship and relate them to the science studies concept of research infrastructures, to eventually arrive at a framework for science platforms. Secondly, we aim to assess how science platforms may affect scholarly practice. To this end, we relate common platform practices to scientific practice. Thirdly, we aim to assess to what extent science is platformised and how this interferes with scientific understandings of quality and autonomy.

At the end of this article, we argue that the potential benefits of platform infrastructure for academic pursuits cannot be ignored, but the commercialization of the infrastructure for scholarly communication is a cause for concern. Studying the effect of platformisation and developing a well-informed perspective on its effects on academia can help understand and predict the benefits and risks for the academic community.

In the field of science studies, infrastructures are widely understood as socio-technical systems that match between the routines of work practice, technology, and wider-scale organizational resources. For instance, Bowker and Star (1999) argue that infrastructures are intertwined with other structures of social arrangements and technologies, and support communities of practice (cf. Bowker & Star, 1998). In that line, Edwards et al. (2013) describes infrastructures as ecologies or complex adaptive systems that incorporate technological standards, social practices, and norms. Blanke and Hedges (2013) further explore the bottom-up nature of infrastructures, arguing that this understanding is crucial for an infrastructure to meet the needs of its users adequately. Hanseth et al. (1996) propose that infrastructures rely on a degree of standardization and compatibility if they are to function effectively (see also Larkin, 2013).

Given this context, one can assert that digital platforms serve as essential research infrastructures. They embody sociotechnical structures, support communities of practice, and integrate societal norms and values. The key characteristic defining these platforms as infrastructures lies in their intricate social and technical connectivity.

Technical connectivity emphasizes the role of platforms as an architecture that allows for vertical and horizontal integration of services, and ultimately for scaling. Unlike non-digital platforms that utilize a “stable core” and a “variable periphery” to achieve flexibility, digital platforms achieve this through coupled and reprogrammable interfaces between the platform and its complementors (Baldwin & Woodard, 2009). This is facilitated by the use of Application Programming Interfaces (APIs) and Software Development Kits (SDKs), which enable data flow with third parties and the integration of software with platform infrastructures (Bodle, 2011; Helmond et al., 2019; Poell et al., 2019). Such interconnectedness among platforms is what forms the basis for a platform ecosystem to emerge (Srnicek, 2017). As described by Gillespie and Helmond, platforms serve as programmable infrastructures that other software can be built upon and run. In that line, Tiwana (2014) defines digital platforms as “the extensible codebase of a software-based system that provides core functionality shared by apps that interoperate with it, and the interfaces through which they interoperate.”

Social connectivity, on the other hand, emphasizes the role of platforms as intermediaries, linking different sets of users through their social infrastructure. Economists describe digital platforms as “multi-sided markets,” bringing together different groups of users and producers (Rochet & Tirole, 2003). While platforms exist in the offline world, such as train stations, the difference in the digital world is their lack of physical and geographical limitations (European Commission, 2016). As noted by Gillespie (2010), platforms provide the “architecture from which to speak or act”. They are a “programmable digital architecture designed to organize interactions between users” (Van Dijck et al., 2018). In essence, the social connectivity of digital platforms is a key feature that enables them to function as intermediaries, linking different sets of users and thus creating multi-sided markets.

Da Silva Neto and Chiarini (2023) provide an insightful historical overview of scientific digital science platforms and a typology that includes a) associating scientific digital platforms with specific social subsystems, b) phases in the research system, and c) their integration into the research process. We find this third dimension particularly illuminating concerning the notion of platformisation. We argue that it signifies a comprehensive and continuous integration of platforms into the research process, ultimately leading to the execution of specific practices exclusively through the platform infrastructure, making them inherently “platformised”. The mediating function of platforms between diverse social subsystems furthermore presents a salient perspective to examine the political economy of platformisation. However, it is in our view not a decisive factor in classifying digital science platforms, which can be identified as such even if they focus solely on the social organization of scientific pursuits. In our view, the defining features of digital science platforms are their social and technical connectivity and hence their capacity to scale in related areas such as human resources, data volume, technical features and geographical reach.

Our understanding of digital platforms implies that they go beyond providing administrative systems or individual tools to enhance research activity, they create space for technical scalability – that is an increasing volume of stored data, or number of functions generated by users or developed based on user-generated data; and at the same time for social scalability – in increasing the number of users, their diversity and geographical reach. To give an example, we consider disciplinary repositories primarily as a research infrastructure, but not a platform as it does not cover sociotechnical scalability to its full extent – in most cases they do not provide space for interaction and enhancement of their functionality based on user-input. On the other hand, services as offered by “Elsevier tools and databases” can be considered as a commercial platform as they offer instruments to design, amend research activities throughout the research cycle (analysis, adjustment of workflows). Similarly, the EOSC (European open science cloud) can be considered as a platform, because apart from research data exchange, it offers not only space for data storage, but also tools for publication, analysis, broadening and improving its capabilities with the increasing number of users and their contributions.

As sociotechnical systems, digital platforms are characterised by a symbiotic relationship between their technical and social connectivity. The interdependence of these two components is a fundamental aspect of platforms that gives rise to three core practices: 1) scaling, 2) standard-setting, and 3) integration of user data for setting up functionalities. By looking at what platforms do generally, we might better understand what they might do to the social and technical organization of science. In the following parts, we will briefly describe how each practice manifests on digital platforms in general and then introduce some examples from science.

While we use “platform” as a value-free term, “platformisation” implies a change in cultural and social practice that a platform promotes and responds to and furthermore a penetration of platform’s technical extensions into the web and the process in which third parties make their data and architecture “platform-ready” (Helmond, 2015). The aim of this section is to trace and discuss the changes induced by platformisation in science. All three platform practices have important advantages for science, especially in terms of efficiency; on the other hand they also present severe risks.

To begin with, one of the consequences of platformisation practices as shown above is a stronger harmonisation or uniformisation of standards. This comes with clear advantages for research: In a fragmented landscape of science-related digital infrastructure, more uniform standards have the potential to promote transparent governance, efficient communication, and the smooth flow of data, rendering research overall more convenient and efficient (see e.g. Schonfeld, 2018). On the other hand, the need for compliance with platform standards can have negative repercussions for epistemic diversity. Strict adherence to design, layout, workflow, business models, and pricing structures may not always align with user interests (Helmond, 2015). The drive for uniformity can hinder innovation, limit the variety of discovery systems and research infrastructures, and constrain the development of a biblio-diverse ecosystem (de Reuver et al., 2018). Therefore, it is essential to balance the benefits of standardisation with the need for diversity to enable platform users to achieve their full potential.

The use of user data to improve services and make them more convenient is one of the core features of digital platforms and indeed a common explanation for their success. It establishes the basis for the personalisation and algorithmic curation of internet searches and content suggestions. This has the advantage of making research faster and more convenient, thus increasing efficiency, for example through individualised reading suggestions based on search histories (see also DFG, 2021 at 4). Despite its clear benefits regarding the usability of services, the use of user data remains a highly contested area in the study of digital platforms. The turn of publishers to the data analytics business has made it clear that science is “no special case on the internet” (Reuter & Söllner, 2022). Criticism of the practice of “surveillance publishing” (Pooley, 2022) concerns privacy rights, informational self-determination, and academic freedom, for the massive collection and analysis of user (meta)data may facilitate science espionage as well as state surveillance of researchers in authoritarian states and beyond (DFG, 2021). Another much-discussed concern in relation to the use of data in general is the perpetuation of existing biases and inequalities (Wachter-Boettcher, 2017). The issue of filter bubbles and echo chambers and their potentially detrimental effect for freedom of opinion and expression as a cornerstone of democracy has extensively been discussed in the context of media (Khan, 2022; more differentiated Bruns, 2019).

The ability of platforms to scale has obvious advantages: Scalable platforms have great potential for scientific research, particularly in terms of speed, accessibility, and discoverability of results. Furthermore, as the above-described example of AmeliCA shows, platformisation bears the potential to work against the fragmentation of scholarly infrastructure and build a “supercontinent of scholarly publishing” (Schonfeld, 2018). For individual users, horizontal integration, leading to services that are “conveniently bundled” (DFG, 2021), could make research more efficient. For those running and maintaining research infrastructure, one can imagine that the net effect of the use of multi-level and multi-service publishing platforms is a significant reduction of research, opportunity, development, and distribution costs. Given the centralised structure and interoperability of these platforms, discovering, indexing and distributing content is made possible with little to no extra work by individual journal publishers (Wrzesinski et al., 2021). At the same time, resources for new and emerging publishing technologies can be pooled and more innovative solutions be developed – examples such as a frequently requested single source publishing suite (Borchert & Hoffmann, 2022) or a smooth integration of collaborative editing and review in existing editorial management systems are proof of that. In the for-profit area, however, the ability of platforms to scale raises significant concerns about an emerging “knowledge industry” (Burgelman, 2021) and the perpetuation of the “oligopoly” of a handful of powerful players (Larivière et al., 2015). Chen et al. (2019) show that globally leading publishers, especially Elsevier, already acquired research infrastructure all along the research cycle. This danger has also been recognised in the latest version of the Budapest Open Access Initiative which recommends the use of research infrastructure “that minimises the risk of (…) corporate control.” (Budapest Open Access initiative [BOAI], 2022).

It can be assumed that platforms, as a special form of infrastructure, will continue to penetrate the practice of research, and therefore it is necessary to address the implications of this platformisation for the normative structure of science. Our assumption is that platformisation fundamentally affects the normative structure of science, and that there is a certain ambiguity inherent in the question of whether these implications are good or bad. A good reference point is the Mertonian principles, which provide a normative structure of science as a system. (Merton, 1973) Applied to the platform context, the five principles have certain ethical implications.

In conclusion, while platformisation may bring benefits such as increased accessibility and transparency for scientific knowledge and speed of discovery, it also poses significant challenges to the values and autonomy of science. It is crucial for science policy to take a proactive approach in regulating platforms and investing in public interest infrastructure to safeguard the integrity of scientific research. By doing so, we can ensure that the potential of platformisation to enhance scientific progress is realised while avoiding its pitfalls.