ChatGPT vs. human-made argumentative essays: Which is more critical?

The integration of artificial intelligence (AI) into the education sector is becoming increasingly unavoidable; however, the ability of AI systems, such as ChatGPT, to generate critically argumentative texts remains insufficiently examined. This study explored whether argumentative essays authored by humans exhibit greater critical depth than those generated by ChatGPT. This study employed a sequential explanatory mixed-methods design. Ten Indonesian university students voluntarily composed two argumentative essays—on the topics of “city dumps” and “slum neighborhoods”—following an experiential learning model, resulting in 20 human-produced texts. ChatGPT was tasked with generating 20 essays on the same topics. The data were analyzed using the Holistic Critical Thinking Scoring Rubric (HCTSR) developed by Facione (2020), as follows: three blind raters assessed all 40 essays for critical thinking using standardized rubrics. Quantitative analysis (t-tests) and qualitative content inspection revealed that human essays consistently achieved higher scores for experiential breadth, perspective diversity, counterargument integration, and emotional resonance. These findings indicate that despite advancements in natural language generation, AI-produced argumentative writing still lacks the critical sophistication that human authors demonstrate. This study indicates that future AI development must move beyond surface-level fluency to better emulate the nuanced socio-contextual sophistication of human reasoning.