Assoc Prof Lim urges education reforms as AI threatens Singapore’s edge

Associate Professor Jamus Lim told Parliament on 25 February 2026 that Singapore’s public education system remains among the world’s best, but warned its strengths may not endure in an artificial intelligence (AI) age that is “rapidly upending” traditional educational advantages.

Speaking during the Budget 2026 debate, the Workers’ Party MP for Sengkang GRC said the Republic’s strong international standings, including top performance in the OECD’s PISA rankings and high placement in the World Bank’s Human Capital Index, should not breed complacency.

Assoc Prof Lim said the challenge was that Singapore had “done exceedingly well till now”, but it was “highly uncertain” whether it would continue to do so as AI advances disrupt how learning, work, and skills are valued.

He cited Prime Minister Lawrence Wong’s view that Singapore’s response to AI “cannot be one of fear”, and said Budget 2026 showed an “explicit tilt” towards helping the country confront AI-related challenges.

Assoc Prof Lim added that he had spoken about AI in his Budget response two years earlier and agreed with the direction, stating plainly that he supported the Budget.

However, he said it would be “foolish” to ignore AI’s transformative potential to “completely overturn” the competitive advantages that underpinned Singapore’s model, arguing that the country could not proceed with “business as usual”.

The changes required were difficult precisely because the education system was not “failing to deliver today”, he said, but was “not adapted to deliver tomorrow”, requiring “foresight, wisdom, and courage”.

Tracing the development of public education, Assoc Prof Lim said mass education began in earnest about a decade before independence, following the establishment of the Ministry of Education, which sought to reconcile diverse approaches in a multiracial society while orienting schooling towards nation-building and workforce preparedness.

He described the effort as “immensely successful”, pointing to schooling attainment rising from fewer than four years in 1960 to more than 12 years by 2015, alongside a science-and-technical emphasis and a system of sorting and tracking that supported economic progress.

Yet he argued the global education landscape was changing rapidly, with AI eroding the “slight edge in knowledge and ability” that schooling previously conferred.

He cited research suggesting AI can compress skills gaps by raising the productivity of inexperienced and lesser-trained workers, including among college-educated professionals performing cognitive tasks, even if aggregate productivity effects remain debated.

In his own teaching, Assoc Prof Lim said AI’s pace had been “breathtaking”, with tasks that previously required “determined and curated prompting” now answered routinely with “speed and accuracy”.

He said he struggled to separate students based on submitted assignments alone, and argued that the traditional pathway of human capital development into “competent, reliable, and trustworthy executors and operators” would no longer distinguish workers.

From this premise, he proposed reforms across primary, secondary, and tertiary education, framing the AI revolution as an upheaval affecting “education at all levels”.

At the primary level, Assoc Prof Lim criticised the persistence of classrooms “as many as 40 students”, saying such sizes affect both children’s ability to receive attention and absorb lessons, and teachers’ ability to manage pupils without excessive stress.

He said large classes were “even less acceptable” in an AI age, disputing the idea that AI could deliver effective “class size of one” personalisation in a way that replaces human instruction.

While acknowledging that education technology can help in areas such as language acquisition and problem-solving, he warned the benefits are often offset by drawbacks including reduced attention spans, diminished memory, and weakened critical thinking.

He argued that the “saving grace” is thoughtful, deliberate use of EdTech as a complement rather than a substitute for teachers, which in turn requires a better educator-to-student ratio.

Assoc Prof Lim stressed the importance of early foundations, saying they shape later academic outcomes and broader life outcomes, and warning against a system that leaves children reliant on tuition and “parents-turned-teachers” to make up for classroom shortfalls.

He said much of primary learning may appear repetitive to adults, but is new to children, and that failure to grasp basics or to become excited about knowledge can produce incurious individuals ill-prepared for an AI-centric world where “absorb and regurgitate” approaches will not be enough.

Instead, he urged a holistic primary school environment as a safe space to ask questions, challenge assumptions, and pursue interests ranging from art and music to sport, alongside languages, mathematics, and science.

At the secondary level, Assoc Prof Lim argued that major high-stakes examinations would become less relevant in an AI age and called for a shift in mindset from testing as a signal of competence and a streaming tool, towards assessment that checks progress and identifies students who have fallen behind.

He noted that education leaders have acknowledged exams are not everything, and cited student perspectives suggesting freedom from an exam-oriented curriculum can make studying joyful and meaningful.

Although continual assessment has long existed, he said weightings historically paid “lip service” to being continual, with midyear and final exams once comprising as much as four-fifths of total grades.

He said the abolition of midyear exams three years earlier would remain “halfhearted” without rebalancing exam weight versus other assessment forms, and warned that a high-stakes year-end exam could leave students feeling unprepared without a midterm gauge.

He added that anecdotal evidence suggested tuition centres were responding by administering mock midterms for students.

Assoc Prof Lim also singled out the PSLE and GCE “O” Levels as “major, consequential tools” for tracking, reiterating calls for reconsideration of these milestones and for expanding through-train options for students who want them.

In an AI age, he argued, “late bloomers” would have more abundant opportunities to catch up, making flexibility more important.

But he said incentives remained key: until the rewards for exam performance change, it is hard to expect families and students to relinquish their exam focus.

He argued that genuine continual assessment improves learning, citing evidence that frequent low-stakes quizzes can strengthen retention and understanding, and pointing to systems that rely on formative assessment and continuous feedback while producing strong student outcomes.

He also warned that early tracking can increase educational inequality without clear achievement gains, noting that Singapore begins subject-based banding and the PSLE around the age range identified in the literature.

While acknowledging MOE’s retention of mixed-form classes within schools, he said sorting between schools remained significant and could undermine tracking benefits while maintaining exam-based sorting pressures.

Assoc Prof Lim stressed that shifting emphasis away from exams did not mean lowering standards, but rather required steady engagement to keep standards consistently high.

He connected this to what he described as a principal AI-era classroom challenge: students who struggle to sustain attention for basic tasks expected of educated individuals, including reading a full book, comprehending accurately, writing with originality, and internalising mathematical skills.

At the tertiary level, Assoc Prof Lim said AI would likely have the greatest impact, declaring his interest as a university academic and arguing that both pillars of university education—high-quality teaching for workforce skills and cutting-edge research for innovation—would be “upended” by AI.

He questioned what it means to transmit knowledge when information is accessible “at the touch of a button”, quickly devalued after graduation, and often obtainable more cheaply, and what it means to set assignments when answers can be produced within minutes through prompts.

He said generative AI can already synthesise and deliver information that many undergraduates struggle to absorb, potentially reducing the need for entry-level positions, while “agentic AI” can execute tasks new graduates previously filled, weakening firms’ incentives to hire them.

Assoc Prof Lim argued that agent-based models, rather than widely discussed foundational models, would drive disruption by producing faster, often more accurate, higher-quality output with minimal guidance.

He added that AI’s reach extends beyond research and analysis to a wide range of professions, naming areas from vacation planning and financial planning to design, data and legal analysis, software debugging, and medical diagnosis.

He cautioned against complacency based on the overall unemployment rate, saying displacement may not yet be captured and that the “future is better inferred” from unemployment among new workers, adding that youth unemployment had trended upwards since mid-2024 and reports of graduates struggling to find jobs were increasing.

Given these shifts, he said tertiary education should move away from rote learning and solving known problems towards cultivating human strengths, including empathy, networking, judgment, creativity, and vision.

He said he already observed in his classes that stronger students are distinguished less by written submissions and more by inquisitiveness, willingness to challenge prevailing wisdom, ability to think beyond conventional constraints, and capacity to bring others together to complete shared tasks.

Assoc Prof Lim also called for revisiting university entry systems that hinge on relative performance in high-stakes, academically oriented exams, noting that he had previously proposed moving towards admissions based on absolute entry criteria, which he argued was now more urgent.

On research, he welcomed Prime Minister Wong’s US$37 billion commitment to the Research, Innovation and Enterprise plan, RIE2030, but argued that funding alone does not produce scholarship, especially in an AI age when research capacity gaps may narrow.

He said what would remain relevant is the sophistication of research questions and approaches to problem-solving, which still depend on what is “unescapably human”.

Assoc Prof Lim argued that the most groundbreaking innovations rely on human imagination and flourish best where societies are open, liberal, and democratic, linking such conditions to human capital accumulation, scientific output, productivity, and long-run economic performance.

He urged autonomous universities to stop “chasing global rankings” for their own sake, and to focus instead on the genuine nature of academic enterprise, expressing confidence that recognition would follow from substance.

Concluding, Assoc Prof Lim said that no matter how education is revamped, there will always be students at risk of being left behind due to natural distributions of ability.

But he argued the AI moment also creates opportunities: teachers can use AI-generated resources to deliver more interesting and customised education, and weaker students may have a more realistic chance than ever before to shine.

He said AI would both compress the talent gap and devalue traditional degrees, opening doors for people with previously undervalued skills to secure a place in the modern economy.

He highlighted those who may be less “book-smart” but gifted in other ways, including hand-eye coordination, social ability, practical intelligence, and a “tinkerer’s mentality”.

Singapore has long embraced ethical, cultural, and religious diversity, he said, adding that it should also embrace neurological and intellectual diversity, beginning with education, and allow diversity to become a “core strength” in an AI age.