Published on 11 Dec 2023

Commentary: Generative AI tools can help, but personalised learning still needs human teachers

Since the launch of ChatGPT nearly a year ago, there has been a surge of exploration and experimentation with this generative artificial intelligence (AI) tool, especially in its application to the field of education.

The prevalent narrative suggests that ChatGPT will transform education with its ability to provide personalised learning. Ask ChatGPT what it can do for education, and the words “personalised learning” keep popping up.

We followed up and asked ChatGPT to explain personalised learning, and it says that it refers to an educational approach where the pace of learning and the instructional approach are tailored to the needs of individual learners.

Indeed, personalised learning often involves the customisation of learning objectives and content, based on a learner's strengths, needs, skills, and interests, with the aim of fostering more effective and efficient learning.

A classic study by American educational psychologist Benjamin Bloom way back in 1984 alludes to the benefits of personalised learning.

It found that one-on-one tutoring can dramatically improve a student's performance, making them do much better than students in a typical large class.

But let’s look at the current state of generative AI’s capability to provide these forms of personalised learning.


It is true that ChatGPT responds to individual queries, thereby giving an impression of personalisation. However, this is a rather surface level of customisation.

Other tools such as search engines like Google and Bing are capable of similar responses, yet we do not credit them with enabling personalised learning.

Advocates of ChatGPT argue that the AI's ability to tailor responses based on the role, context, and target audience specified by the user is an advancement in personalised learning.

While this feature certainly adds an extra layer of specificity, it still falls short of what we traditionally understand as personalised learning.

Over the past year, there have been extensive discussions surrounding the utilisation of ChatGPT in education, which have also included critical examinations of the potential downsides such as plagiarism and academic integrity concerns.

Practitioners have made various efforts to explore the limits of ChatGPT's capabilities, within its limitations, by employing strategic techniques to elicit useful responses for users.

A novice learner is expected to use ChatGPT to learn by asking a series of questions, which can be a valid approach. However, in educational settings, a tutoring or tutor-guided session is needed to initiate and guide meaningful dialogues with students.

Interestingly, innovative methods have been discovered to craft prompts that enable ChatGPT to actively engage in a dialogue with the user.

This aligns with the educational objective of having a tutor-like figure who can lead learners through a Socratic dialogue, fostering deeper understanding and learning.

For example, a teacher can create this prompt and provide it to her students to feed to ChatGPT: “You are an English language teacher in a Secondary 2 class in Singapore. You ask your students to write a composition. The topic is ‘Good deeds by good people’ narrating the story of a good person who has done things to help others. Now I am your student. Please guide me step by step to write this composition.”

Here is another example of getting ChatGPT to help you to prepare for an interview:

Such creative ways of prompting ChatGPT do provide a limited sense of personalised learning. In this aspect, ChatGPT is like a musical instrument, such as a versatile piano.

As users, we are the skilled musicians who explore the instrument's capabilities by playing different melodies, experimenting with chords, and pushing the boundaries of its sound.

By providing creative prompts and interacting with ChatGPT, we can compose unique and harmonious conversations, just as a musician creates beautiful music using the full range of the piano's keys.


Currently, there is a limit on how much you can draw out of ChatGPT, in its current state, to personalise learning.

We assert that real personalised learning involves a deep understanding of a learner's strengths, weaknesses, and knowledge gaps.

Think of our past teachers who were able to provide us with truly personalised learning experiences. They can build on previous lessons, track our progress, and adjust their teaching methods based on the students’ unique learning difficulties and pace.

Generative AI tools like ChatGPT provide the semblance that it can emulate some of these ways of personalising learning but their inability to retain information across multiple conversations is a significant drawback.

Users expect ChatGPT to remember their past interactions when they resume a conversation or session.

That’s true, but ChatGPT is unable to learn from its past responses or accumulate knowledge from earlier completed or other conversations with the same user. This absence of continuity hampers the pursuit of achieving truly personalised learning.

The holy grail of education technology especially with the help of AI has been to support personalised learning at scale.

Generative tools like ChatGPT have come onto the scene, but with the current limitations of generative AI, we need to continue striving for advancements that will better support true personalised learning.

In a recent Instagram post, Minister for Education Chan Chun Sing emphasised the potential of AI as boost potions for teachers and students.

One notable example is the use of ShortAnsFA, an AI-powered tool that expedites the feedback process for teachers by generating initial drafts of grades and comments, which can then be customised and expanded upon.

By utilising such AI tools, teachers can enhance their work and provide a more elevated level of personalised learning support to their students.

In this collaborative partnership with AI, teachers can also leverage their expertise to curate a diverse array of educational experiences, as not all these experiences need to be dependent on AI tools.


Looi Chee Kit is Emeritus Professor of Education at the National Institute of Education (NIE), Nanyang Technological University, and Chair Professor of the Education University of Hong Kong. Wong Lung Hsiang is Senior Education Research Scientist at the Centre of Research in Practice and Pedagogy, NIE.

Read the original article here.

Source: TODAYOnline © Mediacorp Pte Ltd. All rights reserved.