Investigating Linguistic and Content Accuracy of Artificial Intelligence-Assisted Tools for English Writing: A Case Study of English Popular Press Articles

Abstract

This study investigated the linguistic and content accuracy of AI-assisted tools in summarizing and evaluating popular press articles that report empirical research. A qualitative content analysis design was employed to examine the performance of AI in summarizing and evaluating English-language popular press articles. The findings showed that AI-generated summaries and evaluations have a high level of linguistic and factual accuracy. Although the AI-generated evaluations also exhibited substantial factual accuracy, they lacked sufficient supporting evidence from the original press articles. These findings are consistent with previous research suggesting that AI-assisted tools primarily rely on natural language processing algorithms to enhance linguistic accuracy rather than to produce comprehensive evidence-based judgments. Consequently, the AI-generated evaluations frequently omitted the detailed supporting information required for rigorous evidence-based reporting.