Conceptualising Digital Intermediaries on Digital Platforms

The final panel at this excellent Indicators of Social Cohesion symposium in Hamburg starts with the excellent Jakob Ohme, whose focus is on digital intermediaries in knowledge processes on digital platforms. Such platforms lead to context collapse, a levelling of epistemically hierarchies, and a disintegration of formerly fixed sequences in the knowledge process; through this, for instance, journalism has lost its gatekeeping function and information monopoly, actors have switched roles in the information process, and the amount of unverified information that is circulating has increased substantially.

Public communication flows online are now a dynamic network, which is difficult to model. It involves multi-step information flows with large numbers of participants in communication processes who are engaged in many-to-many communication. The pipeline model with its gatekeepers has given way to a platform order where gatewatchers and fact-checkers engage in journalistic roles, sources and speakers can bypass journalistic distribution processes, and recipients can be active produsers of content and can distribute content widely across digital platforms.

The dyadic relationship between senders and receivers of information is thus increasingly intermediated, and intermediaries are becoming a third element positioned between the two others, providing a service to both; this may objectify, criticise, evaluate, and change the dyadic relationship, for better or for worse. Journalists may still play this intermediary role, but it is now optional rather than essential: politicians can speak directly to their audiences, but journalists can also serve to facilitate or accompany this relationship, for instance. This Digital Intermediary Model – or D(X)IM – can then also be translated into a framework for empirical research.

This might begin with a single, communicative instance as an analytical unit; it is user-centric by analysing digital trace data at the individual level; it observes developments over time; allows for the reversal of source and recipient roles; and leaves open the potential for within- and between-actor changes at the horizontal and vertical level. At the actor level, it is then possible to examine intermediation by evaluating the range of services they provide and assess their impact on communication outcomes; at the message perspective, we might focus on the types, contents, topics, and formats of intermediaries’ messages; and at the network level we may study network actor constellations and interactions and the role of intermediaries in message dissemination.

Every communicative action thus also has the power to shape actors on platforms, or the institutional order on platforms itself; in digital intermediation this is different from direct interactions, as the interplay of two actors affects other participants. This approach can then also be used to study social cohesion: it might evaluate the level of engagement; the presence of constructive and respectful interactions; the changes in network density; and the evidence for shared values and norms. It is then possible to study trajectories of social cohesion over time, examine the intermediaries that contributes to this, and explore the importance of social cohesion on an individual level.

A potential operationalisation of this framework might combine the study of digital trace data with periodic population surveys, for instance; increasingly, it should also consider the role of non-human actors – and especially emerging AI agents – as a distinct category of participants in these processes.