Artificial intelligence (AI) has changed higher education faster than we realise. Today, a student can ask an AI tool to produce an essay, summary, slide deck or even computer code in minutes.
Yet many varsities are still grading these same outputs as proof of learning, inadvertently awarding credentials for work that no longer demonstrates what a student truly understands or can do.
Treating this as a problem of cheating, institutions respond with detection software, stricter rules and tighter invigilation. Such measures may reduce obvious misuse, but they do not solve the deeper issue.
Many assessments were designed for a time when producing text and basic analysis required substantial human effort but those tasks no longer reliably measure learning.
If we keep defending old assessments instead of redesigning them, we risk building a compliance machine around evaluations that are steadily losing their meaning.
This forces us to confront a larger question: what is the university’s unique value in a world where knowledge is instantly accessible? Varsities were built for an era when information was difficult to obtain. Lectures made sense: one expert delivered scarce knowledge to many learners. That model worked when books were limited, journals were slow, and expert explanations were hard to find. That era has ended.
The future university
Students can now access explanations, examples, practice questions and instant feedback at any hour, often at very low cost. This does not mean universities are obsolete. It means their role has shifted.
A university cannot compete with machines on content delivery alone. It can remain valuable by focusing on what AI cannot provide: human guidance, ethical judgement, critical challenge and identity formation.
Thus, the future university should be built around encounters, not mass lectures. The most valuable part of higher education will increasingly be the quality of encounters: moments when students are questioned, coached, corrected, stretched and supported by experienced educators and professionals.
These encounters occur in mentoring sessions, research supervision, clinical and workplace teaching, design studios, labs, community projects and structured debates – spaces where students learn how to think, not just what to know.
This also provides a clearer way to interpret the newly launched Malaysia Higher Education Blueprint (RPTM) 2026–2035, which emphasises producing graduates who are knowledgeable, ethical and globally competitive, guided by human values, well-being and sustainability.
It signals that higher education must prioritise judgement and character, not merely certification. But if Malaysia truly wants graduates shaped by values and responsible thinking, the first reform must begin where universities signal what matters most: assessment.
Assessments and AI
Varsities must stop grading what AI can easily generate and shift instead towards assessments that require students to demonstrate thinking, reasoning and accountability.
These include oral examinations and short vivas, supervised problem-solving, applied case analysis, practical projects, studio-based design work, and portfolios developed over time.
Most importantly, students must be required to defend their decisions. Such assessments do not merely reduce misuse; they measure the right things: clarity of reasoning, judgement under uncertainty, and intellectual responsibility
This is also consistent with the direction of RPTM, which calls for higher education to develop individuals who contribute meaningfully to society and solve real-world problems.
In the AI era, the core skill is judgement, not speed. A competent graduate must be able to recognise when an AI answer is wrong, verify claims using reliable sources, cite evidence accurately, protect confidential data, understand bias and fairness risks, and take responsibility for decisions made using AI-generated outputs.
The way forward
One practical step universities can introduce immediately is the requirement for evidence trails in student work: what sources were used, what was verified, what was uncertain, and what the student decided to accept or reject. This simple practice shifts learning from producing answers to demonstrating accountability.
Varsities must recognise the emotional and cultural impact of AI on students. AI can reduce workload, but it can also create pressure: constant comparison, fear of being replaced, and confusion about what is allowed. If universities respond primarily through policing and suspicion, they will increase stress without improving learning.
A healthier approach is to make expectations explicit: what students may use AI for, what they must disclose, what they must verify, and what remains strictly human responsibility.
When rules are clear and assessments are redesigned, anxiety drops and genuine learning improves.
If Malaysia’s higher education system wants to be globally competitive, upgrading digital infrastructure alone will not be enough. The deeper reform is educational: shifting from lecture-heavy delivery and output-based grading towards encounter-rich learning and judgement-based assessment.
Varsities can no longer be defined by what they deliver, but by what they develop. If institutions assess thinking, accountability and ethical reasoning – rather than polished outputs – AI becomes a tool for deeper learning, not a shortcut around it. The future university won’t be measured by how much content it teaches, but by how well it shapes graduates who can think, decide and lead responsibly.
Prof Dr Abtar Kaur Darshan Singh is a professor of innovative digital learning, director of the Digital Learning Hub, and Unesco chair in Harnessing Innovative Technologies to Enhance Quality Teaching and Learning at Asia Pacific University of Technology & Innovation (APU). A two-time Fulbright scholar, she has over three decades of experience in instructional design, digital learning and educational technology. She leads APU’s AIDE team, advancing ethical and impactful AI in education globally, and advises international bodies including Unesco and the World Bank. The views expressed here are the writer’s own.
