Recommender systems are ubiquitous, shaping our experiences on platforms such as Netflix, Instagram, Spotify, job portals, and news outlets. However, for many academics, these systems often operate as ‘black boxes,’ while industry professionals may lack the resources to pursue innovations beyond immediate operational needs.
This panel presents three case studies of academia-industry collaborations. It examines the approaches adopted, the insights gained into recommender systems, the benefits delivered through these partnerships, and the challenges encountered. These examples underscore the value of combining academic inquiry with practical expertise to advance both the understanding and application of recommender systems.
Panel Topics and Presenters
- Value-driven news recommender systems
Dennis Nguyen & Karin van Es (Utrecht University), Arie-Bart de Vries (DPG) - TikTok algorithms and eating disorders
Joris Veerbeek (Data School) & Eva Hofman (De Groene Amsterdammer) - Responsible algorithms in the world of work and news
TBA
What Attendees Will Learn
The three case studies provide the foundation for a critical discussion on the multifaceted role of recommender systems in the media sector.
- Recommender Systems as „Sociotechnical“ Interventions. A complex interplay of professional values, organizational goals, and technical potential shapes the implementation of algorithms for media content distribution. An important question is how different stakeholders influence the design and functionality of these systems. The panel will shed light on these complex processes and underline the importance of inter-stakeholder dialogue.
- Recommender Systems as Subjects of Investigative Journalism. Automated content curation, often geared toward personalisation, has manifold effects. Some of these effects can be considered harmful, making recommender systems a crucial topic for public debate. The panel will thus also explore how journalists can investigate, unpack, and critically report on these systems’ inner workings and societal impacts—a challenge that requires transdisciplinary collaboration.
- Recommender Systems as Ethical Challenges. While algorithmic systems can offer the potential to enrich users’ personal experiences, they also raise critical concerns about inclusion, reach, visibility, societal integration, isolation, and fragmentation. Considering their diverse impacts on different social groups, it is essential to address the ethical challenges these systems pose. This is not merely a philosophical issue but has practical implications for the design and implementation of such technologies.
Transdisciplinary Collaboration between Research & Practice
The panel emphasizes the importance of transdisciplinary cooperation among academics, journalists, media professionals, data scientists, and engineers. It showcases successful collaborations while also addressing the challenges of finding a common approach. This includes both epistemological considerations and practical-logistical challenges.