What question did this study set out to answer?

December 11, 2025Nature CommunicationsOpen Access

Linguistic features of AI mis/disinformation and the detection limits of LLMs

View Full Paper

Ask AI

Bookmark

This study aims to examine the linguistic features of AI mis/disinformation and the detection limits of large language models (LLMs).
Compiled datasets of Chinese AI mis/disinformation from multi-lingual models.
Conducted psycholinguistic and computational linguistic analyses.
Investigated eight language features, toxicity scores, and syntactic dependency distance.
Found implicit linguistic distinctions affecting detection capability of LLMs.
Showed limitations in LLMs' intrinsic detection ability.
Highlighted challenges in applying LLMs for effective information governance.

No takes yet. Share an insight, caveat, or question.

Analysis reveals limits of LLMs in detecting AI mis/disinformation, highlighting challenges in information governance.

Ma et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark