Human Writer
Research on human writing is increasingly focused on comparing and contrasting human-authored text with that generated by large language models (LLMs), particularly exploring the nuances of creativity, fluency, and coherence across different model sizes and training methods. Current studies utilize various NLP techniques and statistical analyses to identify distinguishing features between human and AI-generated text, including vocabulary diversity, narrative structure, and even the detectability of AI authorship. This research is crucial for understanding the capabilities and limitations of LLMs in creative writing, informing ethical considerations surrounding AI-assisted writing, and developing effective methods for detecting AI-generated content in various contexts, such as academic publishing and online security.
Papers
Assessing AI vs Human-Authored Spear Phishing SMS Attacks: An Empirical Study Using the TRAPD Method
Jerson Francia, Derek Hansen, Ben Schooley, Matthew Taylor, Shydra Murray, Greg Snow
CollabStory: Multi-LLM Collaborative Story Generation and Authorship Analysis
Saranya Venkatraman, Nafis Irtiza Tripto, Dongwon Lee