Algorithmic systems, human agency and the future of platform research

As the volume of online content grows exponentially, algorithmic curation has become necessary for delivering content to billions of users daily, such as via recommendation and ranking algorithms. However, there are widespread concerns that these algorithms amplify problematic content or prioritize content that reinforces users' existing beliefs. Although legitimate, these concerns have overshadowed the role that users' own preferences play in determining these same outcomes. Here we argue for a more balanced approach to understanding online content consumption and engagement that considers both algorithmic mechanisms and human agency. We review the existing literature on the role of algorithmic recommendation and users' self-selection in the patterns of users' online diets and show that evidence linking algorithmic curation to problematic outcomes is inconclusive. We highlight the gaps in the existing literature and call for the need for more large-scale data-driven empirical research and experimentation, with the goal of refining the debate on the societal impact of digital platforms and providing a path for future research.