Mitigating the anxiety emotion on consuming personalised feed in Chinese social media

Abstract

Algorithmic social media feeds curate personalised content, often exposing users to anxiety-inducing posts. While anxiety is typically seen as a negative outcome, my research reveals that many users on RedNote (Xiaohongshu) continue engaging with such content for self-improvement and preparatory coping. However, excessive exposure can disrupt well-being, and existing feed control mechanisms are often ineffective or difficult to navigate. My PhD explores design interventions to help users better understand and manage their anxiety in algorithmic feeds. I investigate LLM-powered anxiety trigger awareness tools, self-tracking visualisations, and tangible reflection devices to support self-awareness, emotional regulation, and more intentional engagement. Through empirical studies and user-centered design, I aim to foster healthier interactions with algorithmic feeds. At ECSCW, I seek feedback on my design decisions and ethical considerations, particularly regarding moderate vs. radical interventions in anxiety-inducing content consumption and the ethical implications of applying LLM technology to non-participant user-generated content.