Skip to main content
Article
100%
HD

A recent study tested ChatGPT, the AI‑powered virtual assistant, to see how it affects university students' critical thinking. Forty students from a southwestern U.S. university took part. After completing a survey, six shared their ChatGPT dialogues and two joined in‑depth interviews.

Survey results show the AI boosted the “trigger event,” “exploration,” and “integration” phases (averages 3.6‑3.7 out of 5), but had little effect on the final “resolution” stage. Dialogue analysis uncovered two usage modes: a passive, AI‑directed approach and a collaborative, AI‑supported interaction.

Interviews revealed that, with proper guidance, students completed all four cognitive presence phases, enhancing their critical thinking. In practice, ChatGPT acts as a “more knowledgeable other,” prompting reflection rather than replacing it.

The authors conclude that AI’s real value in education hinges on the quality of human‑AI interaction. Future teaching should aim for co‑participation, where human and artificial intelligence blend to open new epistemic horizons, while keeping student oversight and critical reflection at the core.

Bonus
100%
HD

The study adopts the Community of Inquiry (CoI) framework, which divides cognitive presence into four phases: trigger event, exploration, integration, and resolution. A mixed‑methods design combined quantitative Likert‑scale surveys with qualitative thematic analysis of scripts and interviews. Findings show that AI‑supported interaction enhances metacognition, prompting students to ask deeper questions and link disparate concepts. Yet, the resolution phase remains fragile because it demands autonomous synthesis and critical evaluation—skills the AI cannot yet replace. Neurocognitively, collaborative use of ChatGPT may activate prefrontal circuits linked to abstract reasoning, whereas passive use reinforces more automatic neural pathways. To maximize benefits, educators should design scaffolded activities, provide formative feedback, and embed individual reflection moments, ensuring AI acts as a catalyst rather than a substitute. Developing assessment rubrics that value originality, argumentation, and synthesis is also essential to prevent over‑reliance on pre‑generated answers. In short, effective ChatGPT integration requires a dynamic balance between automation and human oversight, focusing on fostering cognitive autonomy.

Ask AI

Explore key aspects of this story with 3 quick insights

Source

Authors:
Nesma Ragab Nasr
Date Published:
Original Article www.mdpi.com
Fact-check verified on 24/06/2026
Read our Editorial Guidelines & AI Ethics Policy
Creative Commons License CC BY-NC-SA 4.0

Keep exploring

Discover the full archive of science bites or head back to home

Riconnessione al server... Aggiornamento della sessione in corso
Connessione interrotta Impossibile stabilire il collegamento
Sessione aggiornata Ricarica per visualizzare i contenuti