Do We Need Subsidiarity in Software?

Paper Title

Do We Need Subsidiarity in Software?

Publication Info

• Topic area: Ethical considerations in data privacy and control in software design.
• Keywords: subsidiarity, data privacy, user autonomy, surveillance capitalism, chat applications, social media, operating systems, browsers, ethical technology design.

Background and Problem

• Problem / challenge: Contemporary software often centralizes control at the platform level, limiting user autonomy and privacy. This centralization is driven by surveillance capitalism and convenience but raises ethical concerns about data collection and user control.
• Significance: Balancing user autonomy with necessary higher-level interventions is critical for fostering human dignity and addressing privacy concerns in the digital age.
• Motivation and related work: Previous research has explored subsidiarity in social media and open-source AI but has not applied it to broader software contexts. Studies on data tracking and user perceptions of privacy highlight gaps in user control and ethical software design.

Solution

• Proposed approach: Application of the principle of subsidiarity to software design, emphasizing that control should occur at the lowest level possible and the highest level necessary.
• Novelty:
  1. Introduction of subsidiarity as a critical lens for evaluating software ethics.
  2. Empirical analysis of data flow in operating systems, browsers, email, search engines, social media, and chat applications.
  3. User interviews to assess perceptions of necessary control levels across software types.
  4. Comparative analysis of actual vs. desired levels of control to identify violations of subsidiarity.
• Procedure and key techniques:
  • Data flow monitoring using Wireshark to analyze network transmissions during idle and active states of software.
  • Semi-structured interviews with 16 participants to determine desired levels of control for various software types.
  • Reflexive thematic analysis of interview data to identify motivations for control preferences.
  • Comparison of empirical findings with user preferences to evaluate adherence to subsidiarity.

Results

• Concrete findings:
  • Operating systems: Ubuntu operates at a lower level of control, while Windows operates at a higher level.
  • Browsers: All studied browsers (Chrome, Firefox, Brave) operate at a higher level of control.
  • Email: Proton Mail operates at a lower level of control, while Gmail and University Gmail operate at a higher level.
  • Search engines: Duck Duck Go and Brave Search operate at a lower level of control, while Google Search operates at a higher level.
  • Social media: All studied platforms (Instagram, TikTok, X, Bluesky, YouTube) operate at a higher level of control.
  • Chat applications: Both Slack and Discord operate at a higher level of control, with significant privacy violations in Slack.
• Advantage over baselines:
  • Subsidiarity provides a nuanced framework for evaluating ethical tradeoffs in software design, highlighting gaps in user autonomy and privacy.
• Experiments / evaluation:
  • Data flow monitoring on Windows and Ubuntu operating systems.
  • Analysis of browsers (Chrome, Firefox, Brave) and applications (email, search, social media, chat) during idle and active states.
  • Interviews with 16 participants, including technical and non-technical users, to assess control preferences.
• Limitations and future work:
  • Limited scope of software studied (e.g., exclusion of Apple products, Signal, Telegram).
  • Encryption of network packets limited the ability to determine the exact nature of data transfers.
  • Participant pool was limited to university affiliates, potentially biasing results.
  • Future work should explore marginalized perspectives and cross-cultural differences in privacy attitudes.

Summary

This paper introduces the principle of subsidiarity as a framework for evaluating ethical tradeoffs in software design, focusing on data privacy and user control. Empirical analysis of data flow and user interviews revealed significant violations of subsidiarity, particularly in chat applications, where users unanimously desired user-level control but encountered platform-level control. While some software, such as Ubuntu and Proton Mail, aligns with subsidiarity, others, like Slack and social media platforms, fail to meet user expectations. The findings highlight opportunities for designing more privacy-focused software and fostering trust in community-level governance over corporate control.