Answers are grounded in this paper and cite their sources.
This exploration uncovers links between generative AI outputs, training data quality, and scholarly integrity.
This work examines the intricate connections between Generative AI (GenAI), its training data, and the scholarly record through a data‐driven discourse analysis. Meta's Llama3 was trained using pirated data from the file‐sharing space Library Genesis (LibGen) (Reisner, 2015). We investigate metadata of scientific articles from LibGen presumably used to train Llama3, and consider how retractions and issues identified by Retraction Watch (Retraction Watch Database, 2018) could affect GenAI output. This paper is a preliminary exploration of a complex topic and contributes to the discussion on the effects of training data quality, specifically scientific integrity, on GenAI output and the ethics of its collection and use.
No takes yet. Share an insight, caveat, or question.
Ridenour et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: