Not Just Cells in a Dataset: Imagining a Human-Centered Data Science Approach to Social Recommendation Algorithms (Workshop Paper)
Published in ACM CHI (Workshop), 2022
Social Recommender Algorithms (SRAs) often function on large training datasets, comprised of user-generated content that they acquire from the users’ online activities. However, there are several problematic aspects of SRA design, such as obtaining truly informed consent from users before using their data, and potential intended or unintended harms caused by SRAs. At the heart of such problems lies the dehumanization of users by these algorithms, as they reduce users to data points. This working paper presents ongoing work in developing a framework of recommendations for a Human-Centered Data Science approach to SRA design.