🤖 AI Summary
This study investigates the impact of generative AI—exemplified by ChatGPT—on lexical usage and language evolution. Methodologically, it conducts a systematic, cross-domain comparison between ChatGPT-generated and human-produced texts across lexical breadth, frequency distribution, and diversity, employing standardized metrics including MTLD, HD-D, and VOCD to quantify surface-level lexical richness for the first time. Results reveal that while ChatGPT achieves high lexical coverage, its word frequency distribution is markedly uniform, lacking the bursty occurrence of low-frequency words characteristic of human speech and writing; consequently, its overall lexical diversity is significantly lower than that observed in human corpora. These findings uncover a fundamental divergence in lexical production mechanisms between large language models and humans, providing empirical grounding and methodological innovation for studying AI-driven language change.
