PROMPT Codebook

The PROMPT Codebook is a comprehensive framework for the systematic analysis of disinformation narratives, techniques, and rhetorical strategies across digital media. Developed as part of the PROMPT – AI-Driven Methods for Disinformation Detection and Contextualisation project, the codebook provides a detailed categorisation system encompassing 31 code groups and more than 240 specific codes.

It integrates interdisciplinary insights from media and communication studies, linguistics, rhetoric, and semantics, enabling both human and AI-assisted detection of disinformation patterns. The codebook includes categories for narratives, argumentation, verifiability, facticity, manipulation techniques, rhetorical figures, ideological orientation, and linguistic forms.

Designed for application across multilingual and multimodal datasets, the PROMPT Codebook supports AI-driven analysis using Large Language Models (LLMs) to identify subtle linguistic and rhetorical markers of disinformation. It serves as a foundation for comparative studies across six European countries (Latvia, Lithuania, Estonia, Romania, Italy, and France) and is intended to enhance research, media literacy, and democratic resilience against information manipulation.

The PROMPT Codebook (Short Version) provides a structured framework for analysing disinformation across digital media. It organises 16 main categories and 220 subcategories covering narratives, argumentation patterns, facticity, verifiability, manipulation techniques, rhetorical strategies, ideological cues, and linguistic forms. Developed within the PROMPT project, the codebook integrates insights from media studies, linguistics, and rhetoric to support both human and AI-driven identification of disinformation. Designed for multilingual and multimodal data, it enables systematic detection of linguistic and rhetorical markers across diverse online environments and supports comparative analysis across multiple European contexts.