Automated media such as recommender systems, chatbots as well as voice-activated assistants have greatly influenced the way we produce content, access information and engage with the world around us. By determining, to a certain extent, what we see, how we see it and with whom we see it, these technologies have become a key factor of our relationship to ourselves and to others. New large-scale language models like GPT-3 or text-to-image diffusion models like Dall-E are likely to intensify these reconfigurations. A growing number of public service media are looking to implement these technologies into their services while ensuring the enforcement of their core values as well as the well-being of their audiences. However, the conception of socially responsible technologies cannot be based solely on design principles issued from profit-driven organizations. Providing design guidelines for these public institutions is therefore a major concern that must also be addressed by social sciences and humanities. Unfortunately, design-based approaches remain uncommon in these fields, which makes it difficult to identify reliable methodological protocols. To illustrate how social sciences can contribute to these issues, I will be presenting a design-based research protocol developed for a previous research project on recommender systems. This research aimed to identify how providing users with controls over a recommendation algorithm influenced their media practices. Presenting the results of this experiment will allow us to discuss the role that social sciences and humanities can play in the development of new technologies for public institutions.
Claes, A. (2023). Experimenting with Speculative Interfaces: A Design-Based Approach to the Study of Algorithmic Practices. BAIRAL Research Meeting, University of Tokyo. https://hdl.handle.net/2078.5/103279