The Invisible Hand: Algorithmic Curation Within the Facebook News Feed

Media and academia frequently report on misinformation and polarization within the Facebook News Feed. However, most projects focus on describing the effects of misinformation and polarization rather than explaining their causes. The few which attempt to provide explanations neglect to account for the News Feed’s technical history, resulting in an incomplete context and framing. Both cases pose issues to regulators and engineers seeking to improve the News Feed’s social integrity. One cannot hope to improve the platform through changes in policies, algorithms, or regulations without understanding what the problems are and how they arose. Neglecting the causes risks repeating the effects.

This project aims to provide a wholistic account of the News Feed’s development using the recently leaked Facebook Papers and the Technological Momentum framework. Released by the whistleblower Frances Haugen, the Facebook Papers recontextualize present research and provide details of the News Feed’s design and implementation. Through its historical and inertial perspective, Technological Momentum brings the News Feed’s full lifecycle into view and helps reveal how initial design decisions cascaded into later iterations of the system. Facebook’s rapid growth in the 2000s led the company to rely on engagement metrics to filter its content. However, the resulting machine-learning models created and perpetuated a curative system which promotes misinformation and increases polarization. After 15 years of continual development and use, Facebook has grown reliant on harmful engagement metrics and algorithmic curation. Understanding this history places one in a better position to form potential solutions as it reveals the architectural flaws arising from the usage of engagement within the News Feed’s current design. Future work must be done to ensure engagement metrics are used safely within algorithmic systems or to move away from the flawed metrics.

School of Engineering and Applied Science

Bachelor of Science in Computer Science

STS Advisor: William Davis