Researchers have developed a framework to define and measure "AI slop," a term used to describe low-quality AI-generated text. Through expert interviews and span-level annotations, they identified interpretable dimensions for assessing slop, finding that judgments can be subjective but correlate with coherence and relevance. While the framework offers insights into text quality, scalable automated methods for detection remain an open challenge. AI
IMPACT Provides a framework for evaluating the quality of AI-generated text, potentially improving detection and preference tasks.
RANK_REASON Academic paper introducing a new framework for measuring AI slop. [lever_c_demoted from research: ic=1 ai=1.0]
- alphaXiv
- arXiv
- CatalyzeX Code Finder for Papers
- Chantal Shaib
- Connected Papers
- CORE Recommender
- DagsHub
- Gotit.pub
- Hugging Face
- Influence Flower
- Litmaps
- natural language processing
- ScienceCast
- scite Smart Citations
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →