Researchers have developed a method to distinguish between generic and non-generic duplicate content in information operations on social media. By analyzing 187,000 tweets from Russian Twitter Information Operations datasets, they found that generic duplicates, often low-information posts, are frequently misidentified as coordination when using embedding-based methods. The study introduces an LLM-assisted protocol for labeling tweets and trains classifiers to differentiate these campaign types, suggesting that focusing on non-generic duplicates can reveal a more accurate and focused coordination structure. AI
IMPACT This research could improve the accuracy of detecting coordinated inauthentic behavior online by refining how duplicate content is analyzed.
RANK_REASON The cluster contains a research paper published on arXiv detailing a new methodology for analyzing social media data. [lever_c_demoted from research: ic=1 ai=0.7]
- alphaXiv
- arXiv
- CatalyzeX
- DagsHub
- Gotit.pub
- Hugging Face
- Influence Flower
- Russian Twitter Information Operations
- ScienceCast
- X
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →