Knowledge Construction: Impact of AI in Higher Education Teaching and Learning

Zahiruddin Fitri Abu Hassan, Senior Lecturer, University of Malaya

Zahiruddin Fitri Abu Hassan is a Senior Lecturer at Universiti Malaya, specializing in building pathology, concrete durability and construction technology. A passionate advocate for e-learning, he explores new digital tools to enhance education. His expertise led to key leadership roles at the Academic Enhancement & Leadership Development Centre (ADeC), including Head of e-learning (2014) and Deputy Director (2020–2022). After serving as Head of Training at ADeC until 2024, he returned full-time to the Faculty of Built Environment. In his free time, he enjoys rowing, volleyball, cycling and commuting by train to reduce his carbon footprint.Through this article, Hassan emphasizes the role of AI in assessment design and grading, providing a balanced approach in which AI enhances education rather than replaces the essential human elements of teaching, feedback, and intellectual development.

AI for Better Assessments and Easier Grading

By now, many academics are already using AI to create assessment questions and some advanced users have figured out how to use it to grade student work efficiently. With large language model (LLM) tools, all types of questions can be generated easily, ranging from simple questions for formative work to complex open-book exam questions. We need basic prompting skills and enough context inside the chat box to frame output however we like.

On the other end, we can create and apply grading rubrics and scoring systems that evaluate written coursework for anything based on grammar and structure or even detect certain contextual arguments made by the students inside their work. AI can do this task consistently, reflecting the promise of fairness – AI doesn’t get tired as the day progresses or is biased by student behavior in the classroom. This objective approach makes grading more equitable and frees up our time chasing the ever-shifting KPI.

"Now, in 2025, with deep research capability in the most popular AI tools, complex and articulated work can be generated almost instantly, raising concern about academic integrity, but educators are left with almost no reliable tool to catch this cheat other than their instinct"

Instead of waiting a long time to receive comments and feedback on their work, students can receive objective and targeted feedback instantly when the educator has incorporated AI-generated feedback into the assessment regime. An educator may also embed an AI tool to scan students’ weekly formative assignments and flag common misconceptions based on the answers returned by the students. From this data, the educator can address these misconceptions in the following contact session. In a way, the feedback loop is made better with AI as the intermediary, potentially boosting student engagement and learning experience.

However, efficiency and accuracy are only a part of the picture. Ken Robinson once said, “The role of a teacher is to facilitate learning.” An assessment is not purely grading and finding errors; it is about understanding how students learn, providing meaningful feedback and other nuances of facilitating the learning process that AI cannot fully replicate.

If educators lean on AI for all assessment duties, students may miss the rich feedback that fuels intellectual growth. An AI-generated comment bank, for instance, might correct a math solution or grammar mistake but would struggle to guide a student through the way – the reasoning process or the creative insight behind a truly excellent answer. This limitation reminds us that AI, for all its computational power, is a tool best used in partnership with human educators, not as a replacement.

Student Use of AI and the Risk of Bypassing Learning

We also know that students are using AI not just to help with (and even do) their assessment tasks. LLM made it astonishingly easy for students to get answers instantly and even write essays on whatever topic imaginable. Now, in 2025, with deep research capability in the most popular AI tools, complex and articulated work can be generated almost instantly, raising concerns about academic integrity. Still, educators have virtually no reliable tool to catch this cheat other than their instinct.

For the students, the temptation is straightforward. Why struggle through a difficult essay draft or a complex coding assignment when an AI tool can produce a passable answer in seconds? The student probably would not have thought this would short-circuit the learning process. This is the keyword that everybody in education knows too well. That learning is a process. A student who uses AI to generate their essay or code skips the struggle to understand, the problem solving, the lessons from mistakes, the discussion with their peers, and the reflection, which is the entirety of the learning process.

Students may be confused when they take a pen-and-paper exam or are faced with a real-world problem that requires them to apply the concept they are supposed to learn creatively.

A Pedagogical Look at Knowledge Construction

To understand what’s at stake with AI in higher education, it helps to revisit fundamental learning theories. Constructivist pedagogy, for example, holds that learners construct knowledge actively, building new understanding on the foundation of prior experience. An over-reliance on AI threatens to turn learning into a passive affair. Learning never happens when learning through doing, discussing, and reflecting is removed from the equation.

This mirrors the conversational framework by Diana Laurillard, who argues that learning in formal education is essentially a continuing iterative dialogue between teacher and student. The dialogue that involves cycles of discussion, practice, feedback and reflection helps the teacher and learner articulate and refine their understanding. Simply handing information from a machine to a student without those interactive components is not education – it’s information transfer at best.

Other educational frameworks echo this need for active engagement. Bloom’s taxonomy reminds us that higher-order thinking skills (analysis, synthesis, evaluation, creation) are the pinnacle of learning. Yet these are precisely the skills that a shortcut via AI will likely fail to develop.

We are now facing a reality where educators use AI to create assessment tasks, students use AI to generate their responses and educators then rely on AI to grade AI-generated submissions. This raises fundamental questions about the value of learning—if AI mediates every stage of the assessment process, what role do human cognition, critical thinking and intellectual growth still play?

Educators must rethink assessment strategies to ensure meaningful learning continues. AI should be leveraged to enhance, not replace, the learning process, which fosters deeper inquiry, encourages original thought and requires students to articulate and justify their reasoning. AI is reshaping the educational landscape, no doubt about it. However, how we as educators leverage AI to impact meaningful learning will determine whether AI functions as a tool for intellectual growth or the opposite.

Construction

Leadership Development