Graduates in an AI world must be problem solvers, not armchair thinkers

The Straits Times, Graduates in an AI world must be problem solvers, not armchair thinkers

By Professor Phoon Kok Kwang, SUTD President

Academic mastery no longer suffices. Universities will have to adopt a new approach when AI will likely be able to do everything a university graduate can do – better and faster

Imagine being just 18 years old and outperforming experienced professionals in problem solving, not just in terms of superiority of solutioning, but in the speed of ideation and prototyping as well.

This is exactly what is happening with the astronomical rise of artificial intelligence.

In the last year, a group of first-year students at the Singapore University of Technology and Design (SUTD) have been working on real-world projects for industry partners, completing them in days and weeks, instead of months or even years. This is a sharp contrast to the projects that their seniors undertake with their final-year or capstone projects, which typically take between three and six months to complete.

The difference? The former is focused on fast prototyping and leverages AI heavily in their solutioning.

What is significant is that this group of superstar problem solvers will no longer represent the exception, but the norm going forward.

The goal: To build ‘trilinguals’
Indeed, with the incredible advances in AI technology, many questions have been raised – including the continued relevance of higher learning. If a machine is infinitely smarter than us, why do we need to invest time, effort and money in a university education?

With the rapid development of AI and agentic AI that acts independently, a university education may well become a worthless commodity, sooner rather than later. After all, AI will likely be able to do everything a university graduate can do – better and faster.

But this is based on the expectation that university education won’t change, that it remains rooted in the past, teaching undergraduates everything they need to know in the subjects they study with absolute disregard for the multitude of changes around us. It assumes that students need to be taught everything from the basics to their specialisations in order to become “experts” over their entire adult life.

I am convinced that we have reached a critical inflexion point in the history of learning. Educators like me cannot afford to turn a blind eye to the rapid technological developments around us and continue teaching the way we have been used to for the last few hundred years.

The time has come for us to pivot away from academic mastery and focus instead on skills mastery. Learning must no longer be about how much one can memorise – it must be about how one uses the information available to solve problems with real-world value as an end goal.

Indeed, with the concept of lifelong learning inculcated in us through forward-looking government policy even before the widespread adoption of AI, it begs the question: Do we really need to throw the kitchen sink at our undergraduates even before they start work? Is it really necessary for them to know so much about any one particular subject?

Educators, including myself, are now gripped in debate over how much knowledge our students actually need to excel in the real world, and how universities should equip their students with enough knowledge and foundational AI skills so that they can leverage the boundless capabilities of AI.

Specifically, universities must rethink the delivery of higher education and make the distinction between the mastery of knowledge, the mastery of skills, and the mastery of creation. They need to ask: does the mastery of knowledge still hold the same weight now? Do students still need to be experts in any field to excel in a world where AI can easily provide detailed, well-researched information, quickly and succinctly?

How universities answer these questions will shape what and how they teach. Instead of trying to cover everything, courses may focus on how students use knowledge.

At SUTD, students learn by doing.

A computer science major might spend less time on memorising syntax and more on building systems with AI tools, checking results, and fixing errors. An engineering course might shift from routine problem sets to open-ended design work, where students define the problem, test ideas, and deal with trade-offs. In architecture and sustainable design, students might move beyond studying precedents to prototyping solutions, testing them in real contexts, and refining them through feedback. Knowledge would be taught with a clear purpose and tied to use.

Ultimately, institutes of higher learning must determine how a traditional university education can still equip graduates with the necessary skills to excel in a world where basic intelligence is cheap and abundant. Already, AI agents such as personal assistant OpenClaw, and desktop agents Claude Cowork and the more corporate-oriented NemoClaw are trying to do this for individuals and enterprises.

Set against this backdrop, it is vital that students move away from the academic-heavy degrees of the past and accept that they won’t need to “know everything” before they can do anything useful in the real world.

Instead, they should focus on learning how to work with AI effectively to find solutions quickly to become “bilinguals” in AI and experts in their domain, or “trilinguals” who also understand the importance of design and can fundamentally transform how the work is done to raise speed and quality of outcomes.

In architecture, a monolingual practitioner works linearly, iterating slowly with tools like Revit or AutoCAD. A bilingual architect can work faster using AI mainly to refine output. A trilingual architect has a different organisation of work – deploying AI agents to explore options in parallel and shifting expertise from producing ideas to judging them.

In engineering, monolingual practitioners rely on deep but siloed knowledge. Bilingual engineers use AI for modest productivity gains, but core problem-solving stays the same. Trilingual engineers, however, become system builders – combining design, AI and domain to create tools and automations.

The needed adaptation

Moving away from the traditional concept of a university education could well mean a shift away from lessons to a heavier weightage on out-of-class experiences – including internships and hands-on innovation. Graduates must leave the university as doers rather than mere thinkers because AI will far outclass them as the latter.

In China, this is already happening with the explosion of the “one-person company”, where young entrepreneurs set up businesses without any human staff, leveraging entirely a production line of AI agents who can work 24/7, churning out products and solutions.

Set against this backdrop, universities must now review their entire curriculum to ensure that their undergraduates and postgraduates have sound AI foundations to remain strong economic contributors to society.

Courses must be based on human-centred design – finding solutions that link back directly to society, in the same way Apple introduced ideas like the touchscreen iPhone that catered to intuitive human behaviour and shifted mobile design away from physical keyboards and towards a screen-centric and app-driven experience.

Professors will need to change their mindset about going “deep” into a certain topic and accept that depth of knowledge isn’t the be-all and the end-all in this AI environment. They must learn to look beyond the theories that they have become so accustomed to and focus on applying skills in a way that ensures human needs will always remain of central importance.

And employers accept that the graduate of the future will be a very different one from the past but no less equipped in dealing with the challenges of the new world by working with AI even if they do not master all knowledge.

This is not to say that there should be a complete disregard for the mastery of knowledge. A worker with zero knowledge in engineering will not get anywhere in the field even if he uses AI tools.

A world where humans work alongside AI can accomplish far superior outcomes compared to a purely human workforce. But a lot depends on whether universities, employers and people adapt to this reality.

• Professor Phoon Kok Kwang is president of the Singapore University of Technology and Design, the world’s first Design AI university.