Capture and analyze Grokipedia’s public edit logs for CC-licensed articles

Develop a robust data collection and analysis methodology to capture and analyze the public log of edits that Grok displays on Creative-Commons-licensed Grokipedia articles, enabling systematic measurement of how Grok modifies the corresponding English Wikipedia content.

Background

The authors observe a dichotomy between CC-licensed and non-CC-licensed Grokipedia articles. CC-licensed pages include a user-facing log of edits that Grok applied to the underlying Wikipedia content, while non-CC-licensed pages do not.

They attempted to scrape these edit logs but were unsuccessful, and explicitly note that capturing and analyzing this data is deferred to future work.