InterviewDon't outsource your thinking to AI, NUS president warns

SINGAPORE — National University of Singapore (NUS) has built its global competitiveness not only on research excellence, but also on a strong focus on teaching and preparing students to succeed beyond the classroom.

For NUS, the quality of a university should ultimately be measured by how well it equips students with the knowledge and skills they need after graduation — an emphasis that has placed teaching and employability at the heart of the university’s strategy.

“We take teaching very seriously, and many of these things that we do actually do not count toward the rankings — but yet it’s the mission of our university,” NUS President Tan Eng Chye said in a recent interview with The Korea Times.

He emphasized that research remains an essential mission of a leading university, but global rankings — which tend to place heavy weight on research performance — do not always capture what universities accomplish in the classroom.

For NUS, graduate employability is one of the key measures of whether the university is delivering on its educational mission.

“The employability of our graduates is a very key factor, and that is something the government pays close attention to,” he said.

Tan said NUS takes particular pride in receiving strong assessments from global employers because they provide an indication of whether the university is succeeding in preparing its students for the workplace.

“It shows that the global employers value our graduates highly and they are very much in demand,” he said. “That, in a way, partially fulfills what we hope to achieve in teaching and learning.”

That emphasis on education extends beyond the classroom, with NUS giving students opportunities to put their knowledge into practice through entrepreneurship and innovation.

Through NUS Enterprise, the university translates research and academic ideas into startups and connects them with global markets, helping turn knowledge generated on campus into real-world applications.

Tan described research and innovation as part of a continuous cycle: universities invest resources to generate new ideas and knowledge, while innovation turns some of those discoveries into economic and social value.

The university has built links with innovation hubs around the world, including Silicon Valley, Munich, Tokyo and Shanghai, creating networks through which researchers and entrepreneurs can exchange ideas and gain access to different markets.

Its focus on teaching is also shaping how NUS responds to artificial intelligence (AI), one of the biggest disruptions facing higher education.

Tan said universities have little choice but to embrace AI as it rapidly transforms teaching, research, innovation and university operations. But he cautioned against allowing the technology itself to dictate how education should change.

“Pedagogy should be driving AI rather than AI driving how we respond to it,” he said.

For NUS, the key is to let educational goals determine how AI is used, rather than adopting the technology for its own sake.

At the same time, Tan stressed the importance of building strong foundations before students turn to AI.

“Every field, every discipline, has foundational knowledge and skills which you have to train our students without AI,” he said. “When they are comfortable, then you use AI to further enhance their learning.”

Tan compared the approach to how calculators are introduced in mathematics. Students first learn arithmetic and mathematical reasoning without relying on calculators, establishing the conceptual foundations they need before technology is introduced to help them tackle more advanced problems.

“Students need to build their own disciplinary knowledge and thinking skills first,” he said. “Only then should they use AI to enhance their learning, rather than relying on it simply to provide answers.”

NUS began preparing for that shift well before the arrival of ChatGPT in 2022. Tan said the university started redesigning its undergraduate curriculum in 2016, when advances such as AlphaGo demonstrated the accelerating trajectory of AI. It subsequently made quantitative reasoning and computational thinking compulsory for undergraduates, seeking to give students a stronger understanding of data and algorithmic thinking.

Today, students are taught about AI even before beginning their classes, including how to use the technology — and just as importantly, how not to use it.

“You shouldn’t outsource your thinking,” Tan said.

While AI can increase productivity and deepen learning, Tan warned that excessive dependence on it could deprive students of what he called the “productive struggle” essential to education — the process of wrestling with difficult problems, making mistakes and gradually developing the ability to reason independently.

“One of the fundamental things about education is really to teach our students how to think,” he said.

Tan said assessments must also evolve as AI makes it increasingly easy for students to generate answers, shifting the focus toward how they arrive at them.

“You shouldn’t assess based on answers. You should assess the thinking,” he said. “There’s a big difference.”

NUS is also using AI to expand learning opportunities beyond the classroom.

Tan cited an AI-powered role-playing system used in fields such as psychology, nursing, law and social work. Instead of relying solely on actors to simulate patients or clients in class, students can practice with virtual characters based on different scenarios and receive immediate feedback on their performance.

The system allows students to repeat the exercises at any time, giving each of them more opportunities to practice.

Tan said the initiative shows how AI can enhance learning when it is designed around a clear educational purpose rather than adopted simply for the sake of using new technology.

Ultimately, Tan sees the debate over AI as part of a broader question about what universities are for.

“Universities shape society through the graduates they educate, the knowledge their researchers produce and the innovations they bring into the wider world,” he said. “AI could dramatically increase that impact, but only if institutions remain clear about what they want the technology to accomplish.”