ChatGPT vs Human Writing: A Qualitative and Quantitative Study of IELTS Task 2 Authorship

Abstract

Artificial intelligence tools, such as ChatGPT, have recently become increasingly prevalent in the academic world due to their remarkable capabilities for generating essays, including logical structure, contextual relevance, and advanced vocabulary. Undoubtedly, these tools outperform humans in many aspects; however, despite their language-driven strengths, machine-generated content exhibits apparent shortcomings that make it easily identifiable as machine-authored. This article examines the critical deficiencies and limitations of AI writing, with a focus on the significant qualitative and quantitative differences between human-written and automated texts. It emphasises the originality and emotional colouring of human writing while highlighting the absence of these qualities in AI-generated essays despite their linguistic advantages. These findings may guide educators in helping learners refine their writing skills, identify common errors in human writing, and recognise the unique style that distinguishes it from machine-generated writing. The outcomes can be helpful not only in the academic environment but also in AI development, making the generated texts more closely resemble human writing.