Introduction

Artificial Intelligence (AI) has quietly transformed education in ways I could not have imagined when I first stepped into a classroom as a young English teacher more than two decades ago. As an English lecturer at Kolej Matrikulasi Kejuruteraan Kedah (KMKK), I have witnessed the evolution from traditional, textbook-based classrooms revolving around handwritten essays, dictionaries, and physical whiteboards to dynamic learning environments where AI serves as an omnipresent digital assistant. Today, AI is sitting beside nearly every student, helping them brainstorm ideas, preparepresentations, conduct research, or receive instant feedback on their writing. These technologies have undoubtedly enriched teaching by making information immediate. However, my classroom observations also show that this rapid technological shift brings significant educational responsibilities. While AI has become an invaluable tool, it has also introduced subtle behavioural challenges regarding critical thinking and academic independence that educators can no longer afford to ignore.

Classroom Insights: Integrating AI into Language and Engineering Tasks

To explore how AI interacts with 21st-century skill development, I intentionally integrated specific structural applications into our curriculum, which provided clear qualitative insights into student habits. One experience that convinced me of AI’s educational value occurred during a WE023 persuasive writing lesson focused on sustainable living. In previous semesters, many pre-university engineering students struggled to organize their arguments, frequently resulting in repetitive writing and limited vocabulary. To address this, I demonstrated how Chat GPT could be used responsibly to brainstorm concepts, expand technical vocabulary, and suggest alternative sentence structures. I instructed the students to critically compare Chat GPT-generated suggestions with their own original thoughts before choosing which elements best suited their arguments. This process suggested that when used as a guided partner, Chat GPT can assist students in producing more coherent, persuasive, and structured essays while ensuring they remain the authentic authors of their own work.

However, another classroom scenario highlighted the risks of unverified tool dependency. During our Product Pitching project, where students designed innovative products aligned with the United Nations Sustainable Development Goals (SDGs), I witnessed several groups confidently present outdated environmental statistics that had been generated by Chat GPT without cross-examination. This critical incident served as an immediate reminder that while AI can provide rapid answers, it cannot replace human judgment or data verification. I used that opportunity to pause the curriculum and conduct an impromptu session on source evaluation before accepting them as absolute truth.

On a positive note, AI tools appeared to play a constructive role in mitigating affective barriers, particularly regarding student confidence in speaking activities. Many students chose to rehearse their presentations using Chat GPT that offered instant feedback on pronunciation, grammar, and fluency. Students noted that practicing privately with Chat GPT multiple times prior to standing in front of the class helped reduce public speaking anxiety because the automated feedback felt supportive rather than intimidating.

Methodology and Survey Design

Driven by these classroom observations, my curiosity evolved into a structured inquiry. I designed and administered a diagnostic questionnaire to explore four key areas: students’ self-reported perceptions of Chat GPT in learning, its perceived contribution to 21st-century skills, ethical Chat GPT practices, and student awareness of AI’s environmental footprint.

The survey was administered digitally through the Telegram channels of the respective classes to 147 pre-university engineering students, comprising 87 male students (59.2%) and 60 female students (40.8%). To assess changes in environmental awareness accurately, data regarding computational footprints were collected at two distinct intervals: initially as a baseline before any explicit classroom discussion (Pre-Intervention), and subsequently after introducing targeted reading materials regarding data center resource consumption (Post- Intervention). All results reflect student perceptions and beliefs rather than an objective evaluation of cognitive skill gains.

Results and Discussion

Theme 1: Perceptions of AI as a Learning Partner

The survey findings largely confirmed the trends observed during daily classroom instruction. Pre-university engineering students widely perceive AI as an indispensable learning partner that reduces academic frustration and increases task independence.

Theme 2: Perceived Contributions to 21st-Century Skills

The quantitative data also indicated that students feel AI actively supports the development of core competencies necessary for modern professional environments. Specifically, 94% of the surveyed students believed that engaging with AI prepared them more effectively for their future engineering careers. Furthermore, 89% reported an improvement in their communication skills, 87% perceived an increase in their personal creativity, and 82% felt that AI utilization enhanced their capacity to collaborate during complex group projects. For engineering students preparing to enter a technology-driven workforce, the learners view AI literacy as a core graduate competency.

Theme 3: Critical Thinking and Ethical Verification Gaps

Crucially, the study exposed a clear discrepancy between students’ functional reliance on these tools and their critical evaluation habits. It is vital to separate a student’s perception of skill improvement from objective capability. While 90% of the respondents explicitly acknowledged that AI applications can generate inaccurate or fabricated information, only 79% believed that using AI actually encouraged critical thinking. Furthermore, only 72% of the students stated that they consistently verified AI-generated content before pasting it into their academic assignments. This finding mirrors my observations during the SDG project: students frequently mistake surface-level linguistic fluency for factual accuracy. This confirms that technical AI proficiency must always be paired with information literacy and independent critical analysis.

Theme 4: The Environmental Footprint of AI

The most pronounced shift documented in the study related to environmental sustainability. The data collected before and after the classroom intervention revealed that students are generally blind to the physical infrastructure of cloud computing until it is explicitly introduced into the curriculum.

Prior to participating in the educational session, only 41% of students were aware that running large language models requires energy-intensive data centers that consume immense quantities of electricity and water. Similarly, only 35% understood that widespread AI usage directly drives global carbon emissions. However, after learning about these hidden ecological resource costs, 92% of students indicated a desire to utilize AI tools more selectively and responsibly, while 95% stated that technology firms should prioritize investments in greener AI infrastructure. This significant statistical shift suggests that integrating life-cycle and sustainability thinking into AI literacy programs can successfully influence student behavioral intentions.

Conclusion

Conducting this study has transformed both my students’ perspectives and my own approach to teaching. I have come to realize that responsible AI integration extends far beyond the traditional boundaries of preventing plagiarism or citing software inputs. In the modern educational landscape, digital literacy and environmental responsibility are deeply intertwined. True competence involves asking meaningful questions, verifying factual claims, respecting intellectual property rights, safeguarding personal data, and maintaining an active awareness of the environmental consequences of our digital habits.

After two decades in the classroom, I believe the mission of the educator must evolve. We should strive to design environments where human intelligence and artificial intelligence operate in a balanced harmony, rather than allowing technology to override independent thought. Because our pre-university learners will become the engineers, researchers, and leaders of tomorrow, they must develop more than technical mastery. Our highest aspiration should be to cultivate young professionals who possess the ethical grounding and critical wisdom to know when to use AI, why to use it, and most importantly, when to rely entirely on their own human intellect, creativity, and empathy.

References

Anderson, K., & Safdie, S. (2025, February 27). The environmental impact of artificial intelligence. Greenly. https://greenly.earth/en-gb/leaf-media/data-stories/the-environmental-impact-of-artificial-intelligence

Anderson, K. (2025, September 29). AI efficiency: A Gemini reality check with Greenly. Greenly. https://greenly.earth/en-gb/leaf-media/data-stories/ai-efficiency-a-gemini-reality-check-with-greenly

Miller, C. (2025, December 4). The real environmental footprint of generative AI: What 2025 data tell us. Online Learning Consortium.

https://onlinelearningconsortium.org/olc-insights/2025/12/the-real-environmental-footprint-of-generative-ai/

Written By Madam J.D. Kumuthini Jagabalan