The findings are significant as generative AI tools become increasingly common in higher education. Students now use platforms such as ChatGPT, Gemini, Microsoft Copilot, and Blackbox AI to understand course materials, generate ideas, summarize readings, and support academic assignments. The study shows that AI can support critical thinking when students actively question, verify, compare, and evaluate its outputs rather than simply accepting them.
AI Is Becoming Part of Everyday University Learning
Artificial intelligence has rapidly changed how students access information and complete academic tasks. AI can provide explanations, suggest alternative perspectives, help organize ideas, and offer rapid feedback.
However, the expansion of AI use has also raised concerns about overreliance on technology. Students may receive inaccurate information, become dependent on automated answers, or reduce their own involvement in analysis and problem-solving.
Danny Yonathan and his co-authors examined this issue by separating AI use into several dimensions. Instead of treating AI adoption as a single activity, the researchers analyzed the intensity of AI use, digital literacy, reflective AI use, and passive AI use.
The distinction proved important. According to the findings, the quality of students’ interaction with AI has a stronger connection to critical thinking than simple usage frequency.
Study Involved 250 Students from Five Universities
The research involved 250 undergraduate students from five universities, representing academic fields including education, social sciences and humanities, economics and business, science and technology, and health.
All respondents had used AI applications for academic purposes during the previous six months. Data were collected through questionnaires and a critical thinking test.
The researchers measured four major aspects of AI adoption. These included how frequently students used AI, their digital literacy, their reflective use of AI, and their passive use of AI.
Reflective use included activities such as comparing AI-generated answers with other sources, checking information, asking follow-up questions, evaluating arguments, and improving prompts when the first answer was insufficient.
Passive use referred to receiving or copying AI-generated answers without further examination or analysis.
Students’ critical thinking was assessed through abilities including interpretation, analysis, evaluation, inference, explanation, and self-regulation.
ChatGPT Was the Most Widely Used AI Tool
The study found that AI had become widely integrated into students’ academic activities.
Among the 250 respondents:
- 88.8% used ChatGPT
- 62.8% used Gemini
- 30.4% used Microsoft Copilot
- 20.4% used Blackbox AI
The frequency of use was also high. About 32% of students used AI every day, while 43.6% used AI three to five times per week.
Students primarily used AI to obtain additional explanations of course material, develop initial ideas, summarize readings, check grammar, and support assignment preparation.
The average AI adoption score was 3.84 out of 5, indicating a high level of AI use among respondents. Digital literacy and reflective AI use were also in the high category.
The average critical thinking score was 75.68 out of 100.
Reflective AI Use Produced the Strongest Results
The most important finding was the strong relationship between reflective AI use and critical thinking.
The research model showed that AI adoption intensity, digital literacy, reflective AI use, and passive AI use together explained 41.9% of the variation in students’ critical thinking skills.
Reflective use was the strongest individual predictor.
Students who actively compared AI answers with scientific sources, checked information, asked additional questions, identified weaknesses in AI-generated arguments, and refined prompts achieved higher critical thinking scores.
Digital literacy was the second strongest positive factor, followed by the intensity of AI use.
Passive AI use showed a negative direction, but its effect was not statistically significant.
The difference between students with high and low levels of reflective AI use was particularly striking. Students in the high reflective-use group recorded an average critical thinking score of 82.60, compared with 75.40 among the medium group and 68.90 among the low reflective-use group.
The gap between the highest and lowest groups reached 13.70 points.
This pattern suggests that simply having access to AI is not enough to strengthen students’ thinking skills. Students benefit more when they remain actively involved in evaluating and challenging AI-generated information.
Digital Literacy Helps Students Challenge AI
Digital literacy also played a major role in the study.
Students with stronger digital literacy were better positioned to assess source credibility, recognize possible inaccuracies and bias, understand technological limitations, and use AI-generated information responsibly.
For Yonathan and his colleagues, AI should function as a support for human reasoning rather than a replacement for it.
The researchers’ findings indicate that AI can act as a learning tool when students use it to explore ideas, compare perspectives, test arguments, and seek feedback. However, the same technology can weaken cognitive engagement when users transfer too much of the thinking process to automated systems.
Danny Yonathan of Sekolah Tinggi Teologi Berita Hidup and his research team argue that universities should focus on teaching students how to interact critically with AI, rather than merely deciding whether AI should be allowed or prohibited.
Implications for Universities and Lecturers
The findings carry important implications for higher education institutions.
Universities can integrate AI literacy and digital literacy into teaching and learning activities. Lecturers can encourage students to verify AI-generated information, compare answers from multiple sources, identify errors, and defend their arguments through discussion or presentations.
The researchers also suggest that academic assessment should pay greater attention to the reasoning process, not only the final answer.
Students could be asked to document their prompts, explain how they revised AI-generated content, and reflect on the strengths and weaknesses of the information they received.
Such approaches could help universities reduce the risk of students using AI merely as a shortcut while preserving its potential as a learning support tool.
The study also has relevance for policymakers and university administrators developing AI policies. A policy focused only on restriction may overlook AI’s educational potential, while unrestricted use without guidance could increase the risk of dependence and weak academic engagement.
The evidence instead supports structured, transparent, and responsible integration.
Researchers Call for Further Investigation
The authors note that the study has limitations. Because the research captured data at one point in time, it cannot definitively establish a cause-and-effect relationship between AI use and critical thinking development.
The researchers recommend future longitudinal and experimental studies to examine changes in students’ critical thinking before and after AI-assisted learning is introduced.
Future research could also investigate the role of metacognition, self-regulation, learning motivation, teaching design, academic discipline, and lecturer guidance.
Observations, interviews, interaction records, and prompt-history analysis could provide a more detailed picture of how students actually use AI in academic settings.
Author Profile
Danny Yonathan is the corresponding author and is affiliated with Sekolah Tinggi Teologi Berita Hidup. He collaborated with Moh. Rodli of Universitas Mayjen Sungkono Mojokerto, Lucky Dewanti of Universitas Muhammadiyah Bogor Raya, and Pierre Marcello Lopulalan of Politeknik Pelayaran Banten, Indonesia.
Based on the focus of the study, the research team’s work relates to higher education, artificial intelligence in learning, digital literacy, educational technology, and critical thinking. The academic degrees and more specific individual areas of expertise of the authors were not stated in the manuscript provided.
Source
Article title: The Impact of Artificial Intelligence Adoption on University Students’ Critical Thinking Skills in Higher Education Learning
Authors: Danny Yonathan, Moh. Rodli, Lucky Dewanti, and Pierre Marcello Lopulalan
Affiliations: Sekolah Tinggi Teologi Berita Hidup; Universitas Mayjen Sungkono Mojokerto; Universitas Muhammadiyah Bogor Raya; and Politeknik Pelayaran Banten
Publication year: 2026
Manuscript status: Received 21 June 2026, revised 23 July 2026, and accepted 23 August 2026
0 Komentar