Recommender systems are among the most ubiquitous automated systems in the digital environment. Using data and algorithms to connect users with content and each other, and found in everything from short-stay accommodation platforms to dating apps, news websites and social media, recommenders are a locus for social anxiety, pointed critique, and policy scrutiny -- as well as considerable ongoing research and innovation.
Through a combination of presentations, discussions, and exercises, this full-day workshop explores the technical, regulatory, economic and sociocultural aspects of recommenders, situating them in their historical and industry contexts, and articulating their future prospects.
The workshop is facilitated by an international group of leading researchers from law, communication and media, anthropology, and computer science. Participants new to this topic will gain a fundamental understanding of recommenders’ scope and significance; colleagues already actively engaged in recommender research will have the opportunity to contribute insights from their own work.
The workshop begins with presentations on the histories of recommenders, their contemporary scale and significance, and the broader regulatory environments that are sharpening critical attention on them.
Participants will engage in a group exercise designed to broaden and deepen their understanding of what recommender systems are, some ways of classifying them, and their commonalities and differences across environments and sectors. A series of lightning talks drawing on relevant qualitative, critical and experimental empirical projects follows.
We then discuss the ways that social values might align or conflict with the logics of recommender systems in common use today. An interactive small-group exercise examines the significance and complexity of ‘diversity’ in recommender system design, with a particular focus on music recommendations.
The final session explores the likely future directions and ongoing developments in recommender system design. We conclude with a speculative design experiment, in which participants develop ‘pretotypes’ of designs for more responsible recommender systems.
The workshop involves interactive small group exercises, and so we will need the room set up banquet style. There is no particular upper limit on participants but given this setup we anticipate that room capacity will constrain numbers.
1. Introduction and Overview: Jean Burgess
2. Histories, contexts and definitions - Julian Thomas, Natali Helberger and Jean Burgess
- Definitions, histories and prehistories
- Societal significance and regulation
- Recognising recommenders (group exercise)
3. Case studies - Sanne Vrigenhoek, Jeffrey Chan and Stanislaw Piasecki
- News and public service media
- Music streaming
- Multi-stakeholder, considerate recommenders for digital services
- Industry ethnographies
4. Values - Nick Seaver, Patrik Wikström, and Ariadna Matamoros-Fernández
- From accuracy to diversity, serendipity and commonality
- Dimensions of diversity in a popular streaming music service (group exercise)
5. Futures - Natali Helberger, Julian Thomas and Jean Burgess
- How could recommenders better explain themselves?
- The potential impacts of generative AI
- Pretotype a responsible recommender (group exercise)
Jean Burgess is Professor of Digital Media at QUT and Associate Director of the ARC Centre of Excellence for Automated Decision-Making and Society (ADM+S Centre). She researches and publishes on the histories, cultures and technologies of digital media platforms.
Jeffrey Chan is Associate Professor in the School of Computing Technologies at RMIT University and Associate Investigator in the ADM+S Centre. His research focuses on the intersection of AI, machine learning, recommender systems, fairness and explainability.
Natali Helberger is Distinguished University Professor for law and digital technology, University of Amsterdam, and the director of the AI, Media & Democracy Lab. Natali’s research explores the role of law and regulation in realising public values and user rights in an algorithmic society.
Ariadna Matamoros-Fernández is Senior Lecturer in Digital Media at QUT and Associate Investigator in the ADM+S Centre. Her research focuses on the interplay between user practices and platform design and governance in (re)producing structural inequality.
Stanislaw Piasecki is a Postdoctoral Researcher at the Institute for Information Law (University of Amsterdam). His research focuses on the use of AI in media and journalism from a legal and policy perspective.
Nick Seaver is Assistant Professor of Anthropology and Director of the Program in Science, Technology & Society at Tufts University. He is the author of Computing Taste: Algorithms and the Makers of Music Recommendation and has published on ethnographic methodologies for studying algorithmic systems.
Julian Thomas is Distinguished Professor in the School of Media and Communication at RMIT University and Director of the ADM+S Centre. He has written widely about automation and other topics relating to the pasts and futures of new communications and computing technologies.
Sanne Vrigenhoek is a PhD candidate at the University of Amsterdam and a member of the AI, Media and Democracy Lab. Her work focuses on translating normative notions of diversity into quantifiable concepts that can be incorporated in news recommender system design.
Patrik Wikström is Professor of Communication and Director of the Digital Media Research Centre at QUT, and Associate Investigator in the ADM+S Centre. His research examines how digital technologies shape music cultures and economies.
