Unraveling Entangled Feeds: Rethinking Social Media Design to Enhance User Well-being

Paper Title

Unraveling Entangled Feeds: Rethinking Social Media Design to Enhance User Well-being

Publication Info

• Topic area: Social media design and its impact on mental well-being
• Keywords: Social media, algorithmic curation, mental health, user experience, recommender systems, entanglement, design workshops, folk theories, emotional mismatch, user control

Background and Problem

• Problem / challenge: Algorithmic curation in social media platforms creates emotional and mental well-being challenges for users, particularly those with mental illnesses. Users struggle to understand and control the outcomes of their interactions with these systems.
• Significance: Social media is a critical resource for individuals with mental illness, providing community, support, and information. However, its algorithmic design can exacerbate mental health issues, creating a need for systems that better support emotional well-being.
• Motivation and related work: Prior research highlights both the benefits and risks of social media for mental health, including validation and harmful content propagation. Studies show users feel disempowered by algorithmic feeds, which lack transparency and control mechanisms. This paper builds on these findings by focusing on the experiences of users with mental illness to identify design opportunities.

Solution

• Proposed approach: The authors introduce the framework of entanglement, which explains the disconnect between users’ actions and system responses in algorithmic curation. They propose design interventions to address these issues and support mental well-being.
• Novelty:
  1. Development of the entanglement framework, extending Norman’s action cycle to include emotional dimensions in user interactions.
  2. Identification of six specific entanglement issues affecting execution and evaluation phases in user interactions.
  3. Design recommendations to mitigate entanglement, including contextualized engagement, consumption control, and explicit user input mechanisms.
• Procedure and key techniques:
  • Conducted seven design workshops with 21 participants diagnosed with depression, anxiety, or PTSD.
  • Participants completed two tasks: exploring their experiences with algorithmic curation and prototyping ideal platform designs.
  • Used inductive thematic analysis to identify patterns and develop the entanglement framework.

Results

• Concrete findings:
  • Participants developed folk theories to explain emotional and mental challenges caused by algorithmic feeds, such as emotional mismatch, algorithmic leakage, and exploitation.
  • Six entanglement issues were identified: Hiding, Guessing, Flattening (execution phase), and Dangling, Overloading, Disempowering (evaluation phase).
  • Participants proposed design solutions, including modular feeds, real-time engagement controls, and personalized interventions.
• Advantage over baselines: The entanglement framework provides a new lens for understanding user challenges with algorithmic curation, emphasizing emotional and mental well-being rather than traditional metrics like accuracy or engagement.
• Experiments / evaluation: Workshops included demographic-diverse participants who self-identified as having mental illnesses. Activities were designed to elicit insights into user experiences and generate actionable design ideas.
• Limitations and future work:
  • Participant demographics skewed toward white, female, and college-educated individuals, limiting generalizability.
  • Did not collect detailed diagnostic information due to privacy concerns.
  • Future work should expand the study to include diverse populations and explore the impacts of entanglement in other joined systems, such as generative AI.

Summary

This paper examines how algorithmic curation in social media platforms affects users’ mental well-being, focusing on individuals with depression, anxiety, or PTSD. Through design workshops, the authors identified user-developed folk theories and introduced the entanglement framework to explain disconnects between user actions and system responses. They proposed design interventions, including modular feeds, real-time engagement controls, and personalized safeguards, to mitigate entanglement and support mental health. The findings contribute to theory by extending Norman’s action cycle and offer practical implications for designing emotionally aware algorithmic systems.