Reproduce “Distinguishing AI-Generated and Human-Written Text Through Psycholinguistic Analysis” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Chidimma Opara
Teesside University presents a framework for distinguishing AI-generated from human-written text by integrating stylometric analysis with psycholinguistic theories. This approach links 31 stylometric features to four human cognitive processes, offering an interpretable method for AI text detection that clarifies the fundamental differences in text production.
Key insights
- Human writing is influenced by cognitive load and limited working memory, leading to variations in sentence complexity and length that differ from AI-generated text.
- Metacognition and self-monitoring allow human writers to strategically adjust elements like punctuation and contractions, reflecting choices about formality and tone, which AI systems may not fully replicate.
- Lexical access in humans involves drawing on personal experience and semantic memory, often resulting in more diverse and specific vocabulary compared to AI's reliance on statistical patterns for next-token prediction.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.