Nature paper traces state media effects on LLM outputs

A Nature paper titled 'State media control influences large language models' maps pathways from institutional media sources through commercial datasets to generated outputs. Led by Hannah Waight with coauthors Eddie Yang, Yin Yuan, Solomon Messing, Margaret E. Roberts, Brandon M. Stewart, and Joshua A. Tucker, the study carries DOI 10.1038/s41586-026-10506-7. It was received 25 October 2024 and accepted 8 April 2026, underscoring how training data emerges from existing political institutions rather than neutral sources.

Original post

AN#140@RANDOM_WALKER@B_M_STEWART

BR

Brandon Stewart@B_M_STEWART

1/ New @Nature! We study how powerful institutions shape the information environment for LLMs. Commercial LLM training is opaque, so we trace a path from state-coordinated media -> training data -> model responses.

8:07 AM · May 13, 2026 View on X