Published March 15, 2026 | Version v1

Friction-Based Feed Degradation: A Behavioral Design Intervention for Reducing Compulsive Infinite Scrolling in Social Media Feeds

  • 1. Independent author

Description

Infinite scrolling interfaces used by platforms such as TikTok, Instagram and YouTube have become a dominant mechanism for maximizing user engagement in modern social media.

At the same time, social media platforms have enabled unprecedented opportunities for communication, entrepreneurship, and cultural exchange. They allow individuals to build digital businesses, reach global audiences, and participate in economic activity independent of geographic location.

However, infinite scrolling video feeds have also emerged as one of the most powerful behavioral reinforcement systems in modern consumer software. These platforms increasingly deploy autoplaying vertical video feeds designed to maximize engagement through frictionless navigation and rapid reward cycles.

This paper argues that compulsive scrolling behavior is sustained primarily by the speed and reliability of the reward cycle embedded in infinite feed interfaces.

To address this mechanism, the paper proposes a design intervention called Friction-Based Feed Degradation, a behavioral approach intended to reduce compulsive consumption of infinite scrolling video feeds without blocking access to social media platforms.

Rather than restricting usage directly, the system introduces controlled delays and degraded loading patterns after a small number of videos, thereby disrupting the reward timing that reinforces compulsive scrolling. This approach preserves intentional access to social media while weakening the reinforcement dynamics that sustain addictive consumption patterns.

The proposed framework represents a potential application of humane technology principles and suggests an alternative model for behavioral intervention within attention-driven digital platforms.

 

Files

friction-based-feed-degradation-masini-2026.pdf

Files (123.0 kB)

Name Size Download all
md5:126ef5557be6de80382706575e23af48
123.0 kB Preview Download

Additional details

Dates

Submitted
2026-03-15