"It should help me progress and strengthen my Imaan": Understanding Algorithmic Experiences of Islamic Content on Social Media Recommendation Systems

Paper Title

"'It should help me progress and strengthen my Imaan': Understanding Algorithmic Experiences of Islamic Content on Social Media Recommendation Systems"

Publication Info

• Topic area: Algorithmic experiences with religious content on social media platforms.
• Keywords: Algorithmic experience, Islamic content, social media, recommendation systems, religion-aware algorithms, personalization, authenticity, spiritual growth, user needs, HCI.

Background and Problem

• Problem / challenge: Existing social media algorithms are not designed to handle the nuanced values and sensitivities of religious content, leading to risks such as inappropriate recommendations, misinformation, and spiritual harm.
• Significance: Religious content is increasingly consumed online, and mishandling it can have profound consequences for users’ faith and spiritual practices. Addressing this gap is critical for designing systems that respect and support religious engagement.
• Motivation and related work: Prior HCI research has explored algorithmic experiences (AX) in domains like news, music, and health but has largely overlooked religion. Existing algorithms prioritize engagement and profit, often misaligning with users’ spiritual goals. This study builds on AX literature to explore how algorithms influence engagement with Islamic content and proposes solutions for religion-aware systems.

Solution

• Proposed approach: A conceptual framework and design space for religion-aware algorithms tailored to Islamic content, emphasizing authenticity, appropriateness, and spiritual growth.
• Novelty:
  1. Empirical understanding of how algorithms influence Muslim users’ engagement with Islamic content.
  2. Advocacy for religion-aware algorithms that cater to nuanced religious needs.
  3. Introduction of a hierarchical framework of users’ needs for Islamic content (authenticity, appropriateness, spiritual growth).
  4. Actionable design implications for creating balanced, trustable recommendation systems.
• Procedure and key techniques:
  • Conducted semi-structured interviews and focus groups with 24 Muslim participants from Bangladesh.
  • Explored users’ motivations, experiences, and expectations regarding algorithmic recommendations of Islamic content.
  • Analyzed data using reflexive thematic analysis to identify key themes and insights.

Results

• Concrete findings:
  • Algorithms support spiritual growth by personalizing content and providing timely reminders, but personalization is often shallow and lacks depth.
  • Participants reported risks of inappropriate or unauthentic recommendations, which could harm faith and spiritual practices.
  • Users desired algorithms that prioritize spiritual growth over engagement, with safeguards against misinformation and inappropriate content.
• Advantage over baselines:
  • Proposed religion-aware algorithms address gaps in existing systems by incorporating religious sensitivities, user needs, and expert oversight.
  • Framework prioritizes authenticity and appropriateness, which are often overlooked in traditional recommendation systems.
• Experiments / evaluation:
  • Qualitative study with 24 participants (mean age = 35.1 years, SD = 11.4), including diverse genders, professions, and educational backgrounds.
  • Platforms studied: YouTube (100%), Facebook (79.2%), TikTok (4.2%), Instagram (4.2%).
  • Data collected through interviews (35 minutes) and focus groups (90–110 minutes).
• Limitations and future work:
  • Study focused on Muslim users in Bangladesh; future research could explore other religions and cultural contexts.
  • Did not involve Islamic scholars; future work will include co-design workshops with scholars, users, and designers.
  • Need for longitudinal studies to assess the long-term impact of religion-aware algorithms.

Summary

This study investigates how social media algorithms influence Muslim users’ engagement with Islamic content, revealing both benefits (e.g., spiritual growth) and challenges (e.g., risks of unauthentic or inappropriate recommendations). It introduces a hierarchical framework of user needs—authenticity, appropriateness, and spiritual growth—and proposes a design space for religion-aware algorithms. The findings emphasize the importance of integrating religious sensitivities, user needs, and expert oversight into recommendation systems. Future work will involve Islamic scholars and explore broader religious contexts to refine and implement these systems.