Description

The SMORES project is designed to simulate "The Friendly Algorithm Store" , an ecosystem where users can choose between multiple recommendation algorithms developed by various third-party entities. This project seeks to explore the implications, benefits, and limitations of such a decentralized structure on three key stakeholders: consumers, content providers, and platforms.

By moving away from the traditional monolithic recommender system model—where a single platform dictates recommendation logic—this approach examines how distributing algorithmic control can impact stakeholder utility, competition, and innovation. It provides an experimental framework to assess whether algorithmic plurality leads to improved user experiences, increased fairness for niche providers, and a more balanced distribution of exposure and profit across the ecosystem.

Moreover, SMROES fosters a space for creative algorithmic design, allowing third-party developers to introduce novel recommendation strategies tailored to different user needs. By decentralizing recommendation power, this project contributes to the broader discourse on algorithmic governance, transparency, and user agency in digital ecosystems.

Related Publications

Decoupled Recommender Systems: Exploring Alternative Recommender Ecosystem Designs Anas Buhayh, Elizabeth McKinnie ,and Robin Burke In RecSoGood Workshop at the 18th ACM Conference on Recommender Systems, 2024