<![CDATA[This study investigated the morphological characteristics of AI-generated English texts in comparison with human academic writing, focusing on affixation, compounding, lexical patterns, and word-formation processes. It specifically aimed to identify morphological differences between both text types, examine how AI language models construct academic discourse at the word-formation level, and determine whether morphological analysis can effectively distinguish machine-produced from human-written academic texts. Ten academic texts — five ChatGPT-generated outputs and five drawn from student essays and final assignment articles — were purposively selected based on comparable topic, length, and register. Employing a qualitative descriptive design with a comparative approach, data were collected through documentation and analyzed using qualitative content analysis framework of data condensation, data display, and conclusion drawing. Morphological features were systematically classified into derivational affixes, inflectional affixes, compounds, acronyms, and lexical patterns. The findings revealed notable distinctions between the two text types. AI-generated texts displayed higher frequencies of derivational suffixes (186 instances) and lexical repetition (79 instances), with heavy dependence on standardized suffixes such as -tion, -ment, and -ity to sustain academic formality. Human-written texts, conversely, exhibited greater morphological complexity through multi-affix constructions like misinterpretation and unpredictability, richer compounding patterns including self-regulated learning and cross-cultural communication, and broader lexical diversity achieved through synonym substitution. The study concludes that, while AI-generated writing demonstrates grammatical consistency, human academic discourse remains superior in lexical richness, morphological creativity, and contextual adaptability.]]>
