Data Ethics in the Fediverse: Analyzing the Role of Instance Policies in Mastodon Research
Mareike Lisker, Helena Mihaljević
TL;DR
The paper addresses the mismatch between Mastodon instance level data policies and how researchers collect and use user generated data. It employs a systematic literature review of 29 Mastodon data studies to evaluate the extent of policy engagement and data handling practices. The findings reveal limited engagement with instance policies, licensing and anonymization issues, and transfers of data to external APIs, underscoring ethical complexities in decentralized platforms. The authors propose actionable guidance for researchers, ethics committees, and platform developers, including machine readable policy representations and user level consent mechanisms, to improve ethical research practices in the Fediverse.
Abstract
This article addresses the disconnect between the individual policy documents of Mastodon instances--many of which explicitly prohibit data collection for research purposes--and the actual data handling practices observed in academic research involving Mastodon. We present a systematic analysis of 29 works that used Mastodon as a data source, revealing limited adherence to instance--level policies despite researchers' general awareness of their existence. Our findings underscore the need for broader discussion about ethical obligations in research on alternative, decentralized social media platforms.