Artificial intelligence (AI) is becoming a common learning companion for students, especially in science, technology, engineering and mathematics (STEM). With a few prompts, students can now obtain explanations, solve equations, generate code, summarise readings and structure assignments.

While this offers clear benefits, it also raises an important concern for educators and policymakers: are we strengthening student learning, or slowly weakening independent thinking?

This question deserves attention as Malaysia pushes for a more innovative, digitally capable and future-ready workforce.

STEM education

STEM education is not only about producing graduates who can use advanced tools. It is about developing individuals who can reason carefully, solve unfamiliar problems and make sound judgments in real-world situations.

AI can certainly support that goal when used well. It can make difficult concepts more accessible, provide immediate feedback and help students practise with greater confidence. For weaker learners, it can reduce barriers and improve engagement. For lecturers, it can open up opportunities for more personalised teaching and faster academic support.

Yet the risk begins when AI moves from being a support tool to becoming a substitute for thought. In STEM disciplines, learning is not merely about arriving at the correct answer. It is about understanding how the answer is reached, why one method is chosen over another, what assumptions are involved, and whether the result is reasonable in practice.

If students rely on AI outputs without understanding the logic behind them, they may complete tasks efficiently while learning very little.

Critical thinking

This is why AI dependency can be so deceptive. On the surface, performance may appear to improve. Assignments are completed more quickly, reports look polished and answers sound convincing. But behind this efficiency, students may be doing less analysis, less questioning and less intellectual struggle. Over time, this can weaken critical thinking, resilience and confidence in solving problems independently.

That should concern us because Malaysia’s future challenges will not come with ready-made prompts. Whether in energy, healthcare, infrastructure, manufacturing or cybersecurity, graduates will need to deal with uncertainty, evaluate trade-offs and make decisions in complex contexts. These are capabilities that cannot be built through automation alone.

The response, however, should not be to reject AI. A blanket ban would neither be realistic nor productive. The better approach is to guide students to use AI responsibly. They should be taught to verify outputs, question reasoning, detect errors and judge whether a response makes sense. AI should help students think better, not tempt them to think less.

Institutions also need to review how learning is assessed. If students are rewarded mainly for polished answers and speed, dependence on AI will naturally grow. But if assessment places greater emphasis on explanation, reflection, oral defence and applying knowledge to unfamiliar situations, students will be encouraged to demonstrate genuine understanding.

Smart educational aid

This is an AI-generated image.

As Malaysia continues to advance its digital education agenda, the real issue is not whether students use AI. They will. The more important issue is whether our education system still nurtures the habits of mind that matter most: curiosity, discipline, judgment and independent problem-solving.

AI should remain a smart educational aid. It must not become a silent trap.

Ir Ts Dr Renuga Verayiah is a senior lecturer, researcher and consultant specialising in electrical power system and renewable energy integration for modern power systems at Universiti Tenaga Nasional (UNITEN), Malaysia.