Assessing Privacy in Social Media Aggregators

Gaurav Misra, Jose M. Such, Lauren GillIWPE@SP 2017 · OpenReview

TL;DR

This paper introduces a novel three-step methodology to assess the privacy implications of Social Media Aggregators (SMAs). It reveals a significant disconnect between SMA privacy policies and their actual interface operations, with a substantial portion of analyzed SMAs exhibiting "broken traceability," leaving users with limited transparency and control over their data.

What Problem Does This Paper Address?

In today's hyper-connected world, social media has become deeply integrated into our daily lives, with a significant portion of this engagement happening through mobile devices. Social Media Aggregators (SMAs) have emerged as convenient tools, allowing users to manage multiple social networking site (SNS) accounts from a single application. While offering convenience, SMAs also present a unique privacy challenge: they consolidate data from various SNS accounts and can potentially access sensitive mobile device data (like location, contacts, etc.). This concentration of data raises concerns about user privacy. The paper addresses the critical question of how well these SMAs consider and implement user privacy, specifically focusing on the transparency and control mechanisms they offer. The authors aim to identify gaps in current practices and provide a framework for evaluating SMA privacy.

Methods

The researchers developed a three-step methodology to assess SMA privacy:

  1. Data Permissions Inspection: This involved examining the mobile data permissions requested by SMAs (e.g., access to photos, location, contacts) and the social media data they access when users log in. This was done by analyzing app permission screens, device settings, and by creating social media accounts to authorize SMAs and observe granted permissions.
  2. Privacy Policy Analysis: The study reviewed the privacy policies (or equivalent documents like "Terms of Service") of the SMAs to assess their compliance with app store vetting policies and to understand how they communicate data collection, usage, and purpose.
  3. Traceability Analysis: This qualitative step focused on mapping the privacy implications identified in policies to the actual transparency and control mechanisms offered through the SMA's interface operations. They categorized the traceability into "Complete," "Partial," and "Broken" to quantify the alignment between documented policies and user experience.

Datasets

The study evaluated 13 popular Social Media Aggregators (SMAs) sourced from three app stores: Google Play Store, iTunes, and Cydia. The specific SMAs included were iSocial, Hootsuite, Buffer, Social Networking all in one, Social Media all in one, Everypost, Social Media, Hootsuite (for Twitter & Social), Buffer (for Twitter & Social), Everypost (for Twitter & Social), Social Media Vault, Social butter, and Social hub.

Key Insights and Ideas

  • Data Concentration is a Privacy Risk: SMAs, by design, aggregate data from multiple sources, creating a single point of potential vulnerability.
  • "All or Nothing" Permissions are Problematic: Mobile applications often require broad permissions, and users may not fully understand the implications of granting them.
  • Privacy Policies are Often Insufficient: Even when present, privacy policies can be vague or disconnected from the actual user experience, leading to a lack of transparency.
  • Traceability is Key to Trust: The disconnect between what a privacy policy promises and what the app actually does (or allows the user to do) is a major concern. "Broken traceability" means users are often unaware of data collection or lack control.
  • App Store Vetting Isn't Always Enough: While iTunes and Google Play have vetting processes, the study found SMAs without privacy policies, indicating potential loopholes. Cydia, with no vetting process, also presented privacy concerns.

Key Results

  • Data Access Varies Widely: The study found a significant variation in the types of mobile data accessed by different SMAs.
  • Absence of Privacy Policies: 5 out of the 13 evaluated SMAs did not provide any privacy policies, directly contradicting app store guidelines.
  • Widespread "Broken Traceability": A substantial proportion (45%) of SMAs exhibited "broken traceability," meaning there was a significant disconnect between their privacy documentation and interface operations.
  • Limited Complete Traceability: Only about 19% of the cases showed "complete traceability," indicating a high level of transparency and control.
  • Lack of User Control: Less than a quarter of the analyzed results indicated complete user control over privacy implications.

Limitations and Future Work

The authors acknowledge that the qualitative analysis of privacy policies and interface traceability was less objective and more time-consuming due to inconsistencies in documentation and interface operations.

Future work suggested includes:

  • Extending the Methodology: Applying the developed methodology to other apps that link social media accounts (e.g., gaming apps) to see if similar privacy issues are prevalent.
  • Automation: Developing automated solutions for traceability analysis, potentially by standardizing privacy documentation and defining privacy implications more clearly. This could lead to automatic certification of SMAs.
  • Improving User Awareness: Greater emphasis on educating users about the implications of granted permissions is crucial.
  • Developer Best Practices: Encouraging developers to provide clearer privacy policies and more granular control over data access.

Summary updated

BibTeX
@inproceedings{misra2017assessing,
  author = {Gaurav Misra and Jose M. Such and Lauren Gill},
  title = {{Assessing Privacy in Social Media Aggregators}},
  booktitle = {IWPE@SP},
  year = {2017},
  url = {https://openreview.net/forum?id=9H9mA9M3d8},
}
Figure