*Is AI the answer to bridging the tuition gap?*
https://www.straitstimes.com/opinion/is-ai-the-answer-to-bridging-the-tuition-gap
2025-04-30
By---Professor Looi Chee Kit is Emeritus Professor of Education, National Institute of Education (NIE), Nanyang Technological University (NTU), and research chair professor of the Education University of Hong Kong.
Dr Wong Lung Hsiang is senior education research scientist at the Centre for Research in Pedagogy and Practice at NIE, NTU.
=====
In March 2025, Microsoft co-founder Bill Gates declared that AI or artificial intelligence could replace human teachers in a decade. If that proves true, the first to be replaced may not be mainstream educators but traditional private tuition teachers.
Classroom teaching is far more complex. School teachers do much more than explain content; they manage student behaviour, work with parents, support different learning needs and build trust with their students. These human aspects of teaching are still well beyond what AI can do.
Some forms of private tuition – especially those focused on exam techniques and repetitive practice – follow structured routines that AI can now handle quite well. Notably, the Ministry of Education has also voiced concern over tuition centres that exploit parents’ anxieties and is exploring ways to discourage such commercial tactics.
These developments signal a critical moment to rethink how we define effective academic support and whether AI might offer a healthier alternative.
For many Singaporean families, private tuition has become a default response to academic pressure. Parents who can afford it often sign their children up for multiple subjects, sometimes beginning in primary school. The driving belief is simple: tuition raises grades.
But this comes at a financial and emotional cost. Singapore households spent $1.8 billion in 2023 on private tuition. For lower-income families, this creates a stark dilemma: either stretch limited resources to pay for their children’s tutoring, or risk having them fall behind. This disparity has long fuelled educational inequality, with students from affluent backgrounds gaining access to more academic support.
But what if a new kind of AI – one that goes beyond mere question-answering – could offer students a different path: personalised, affordable and genuinely supportive of deeper learning? That possibility is no longer theoretical.
The AI tutor?
Today’s chatbots, like ChatGPT and DeepSeek, typically respond reactively to prompts, answering one question at a time. We are now seeing growing interest in a different kind of AI – sometimes referred to as “agentic AI” – which can initiate tasks, plan steps, track progress and adapt to the learner over time.
While the underlying ideas are not entirely new, recent advances have made such systems much more feasible, accessible and ready for meaningful application. A striking example of this shift is Manus AI, which recently went viral for its ability to manage complex tasks such as writing, replying and follow-ups with minimal prompts.
Although not designed for educational purposes, this illustrates how far agentic AI has advanced in task coordination, memory utilisation and autonomous planning. These capabilities, if repurposed in education-specific designs, could enable AI tools to not only answer questions but also guide students through structured learning journeys, diagnose misconceptions and personalise feedback.
While some enrichment centres already promise inquiry-based and critical thinking approaches, these remain cost-intensive and variable in quality. AI offers the potential to make such personalised, higher-order learning accessible to a much broader student base.
In a country where education authorities have raised concerns about tuition centres fuelling parental anxiety, such as worrying if their child needs an extra push to do well in important exams, these tools may offer a much-needed shift in mindset and a reframing of out-of-school learning as something that empowers students rather than pressuring them to chase grades.
This represents a contrast to common approaches that may reinforce exam anxiety and competition. In contrast, well-designed AI tools can reframe learning as a process of growth and reflection, offering non-judgmental feedback and custom support that promote intrinsic motivation over grade-chasing.
Some education-focused versions are in development globally, aiming to provide students with on-demand scaffolding, adaptive feedback and step-by-step guidance tailored to each learner’s profile. These tools hint at a future where personalised learning can be democratised – not by replacing teachers but by extending high-quality academic support beyond school hours, even in households that cannot afford private tuition.
More On This Topic
AI helps teachers to work smarter but not ready to take over their jobs
P6 and Sec 4 students flock to tuition centres for mock exams after scrapping of school midterms
Going beyond the answers
Imagine AI-powered agents that guide students through revision in a way that encourages curiosity and builds mastery. These agents would do more than correct grammar or summarise passages; they might track a learner’s misconceptions, adapt feedback to match readiness levels or nudge learners to reflect and self-correct.
Consider a pupil preparing for the PSLE science exam. An AI companion could help break down open-ended questions, highlight key misconceptions such as misunderstanding photosynthesis or energy conversion and guide the learner through systematic revision with tailored feedback.
Instead of overwhelming students with answers, they scaffold understanding – like a wise peer who knows just when to offer a hint or challenge an idea.
Such tools could benefit not just individual learners but also Singapore’s long-running community tuition efforts. Initiatives like the Chinese Development Assistance Council (CDAC) Tuition Programme, Mendaki Tuition Scheme and Sinda Tutorials offer academic support to students from lower-income households.
Many of their volunteer tutors are caring and committed, but may not always have formal training in pedagogy or in-depth content knowledge. AI-powered companions could help bridge that gap by offering volunteers suggestions, prompts or real-time feedback during sessions, acting as effective co-pilots.
Meanwhile, students who miss lessons or cannot attend tutoring for every subject can rely on these tools as “always-on” study partners. These AI companions can pose guiding questions on science worksheets, assist with oral language practice or help students navigate multi-step maths problems – all while tracking progress and providing timely, encouraging feedback.
While the importance of developing AI literacy is often emphasised, it is still worth briefly noting that students should have a basic understanding of how to use these tools effectively and ethically, recognising both their capabilities and limitations as well as integrating them thoughtfully into their learning journey.
Importantly, we do not advocate replacing human mentors or tutors, particularly those who provide the essential emotional support and a deep understanding of each student’s context, which are vital for effective learning.
Rather, agentic AI can extend the reach and effectiveness of human mentors. With thoughtful design and oversight, it could offer a practical and inclusive alternative.
To advance beyond traditional learning models, we must envision new models of learning support. Agentic AI won’t replace teachers who evolve alongside it – instead, it can help shift the perception that costly tuition is the sole route to academic success.
Even as we reimagine the role of AI in private tuition, important challenges remain. Students still need to self-regulate, and not all will thrive in screen-based, self-directed environments.
Some parents may feel added pressure to support learning at home. And while AI can deliver content with speed and precision, it cannot replace the encouragement, intuition or inspiration of a gifted tutor.
As we embrace these new tools, we must also safeguard the human connections that make learning meaningful.
=====
Professor Looi Chee Kit is Emeritus Professor of Education, National Institute of Education (NIE), Nanyang Technological University (NTU), and research chair professor of the Education University of Hong Kong.
Dr Wong Lung Hsiang is senior education research scientist at the Centre for Research in Pedagogy and Practice at NIE, NTU.
=====
More On This Topic
‘Over 800 messages on A-level day’: Enter the emotional support tutor-therapist
Teaching AI in schools needs to go beyond a skills-based approach
Posts
