3 ways Singapore is trialling AI in education

The launch of AI chatbot ChatGPT in November 2022 took the world by storm. While many were intrigued by the possibilities the chatbot presented with active users on the platform crossing a record-breaking 1 million in less than a week, some have raised concerns about what such technology will mean for the future of education. 

In the United States, public schools in New York City and Seattle have already banned access to ChatGPT. A statement by Jenna Lyle, a spokesperson from the New York City Department of Education, expressed concerns that the artificial intelligence tool could have negative impacts on student learning and encourage plagiarism, reported The Guardian. 

But innovation in education is not a technology problem, says Associate Professor Ben Leong, Director of Singapore’s AI Centre for Educational Technologies (AICET). Rather, he believes that it is a pedagogy problem. Leong is also an Associate Professor teaching computer science at NUS, and the Chief Data Officer (Technical) at AI Singapore. 

Indeed, some educators view ChatGPT as a new tool for teaching instead of shunning the technology. 

Some teachers have begun integrating responses from ChatGPT into their curriculum, requiring students to evaluate the responses given by the chatbot to improve their critical thinking skills, according to the New York Times. Likewise in Singapore, an instructor from the National University of Singapore expressed in a radio interview that tools like ChatGPT can even help to improve students’ writing. 

Such possibilities are the very reason why AICET was launched. A research centre hosted by AI Singapore and funded by the Smart Nation and Digital Government Office, AICET works with the Ministry of Education to launch AI projects that will improve the education system locally. 

“Teachers can imagine new ways to teach, but technology is a key enabler that most do not have access to,” Leong says. “We build tools that help teachers connect with more students and teach effectively online, so as to increase their reach to many more people and eventually, across geographical boundaries.” 

Here are some of the ways Leong expects AI will drive innovation in education.

1. Adaptive learning

AI can help provide a more personalised learning experience for students, especially on online learning platforms. 

The problem with normal lesson design is that the pace of learning is often catered to the median, Leong says.This creates a two-fold problem. Slower students will fall behind and lose motivation, while faster students tend to burn out if they speed through the lessons too quickly.

“It’s not about what [students] need to know, but how much time they need to learn it,” Leong explains. This is where AI can come in to manage the timelines for each student. 

AI can automatically detect when a student is taking longer on assignments, and give them more time to understand each module. Meanwhile, the faster students will gain access to each module in shorter intervals. This ensures that lessons are still paced out, preventing them from burning out too quickly.

“We give content at the right place, so that they learn the best at the right pace,” Leong says. 

2. Exam grading

Another way that AI can be integrated into education is in exam grading. 

Traditionally, grading exam scripts is a tedious and menial process. Teachers would have to lug stacks upon stacks of papers home, before going through them one by one. The process is also prone to human error, as mistakes can easily be made when teachers have to manually tally the scores from each individual question. 

This is why AICET has developed the SoftMark programme, which will allow teachers to easily mark on the go. SoftMark works through a type of AI known as Computer Vision. This involves devices being able to recognise and derive meaningful information from images. 

With SoftMark, teachers simply need to scan the exam scripts and the AI is able to determine the fields in which students have inputted their answers. Teachers can then mark as per usual using a tablet, and the AI will tabulate the scores for each individual question as they are being graded. 

The SoftMark programme is already being piloted in St Andrew’s Junior College today, and Leong is looking to add new features like grouping similar answers together for more consistent marking. 

3. Automated feedback system

The third way that Leong hopes AI can be integrated into education is through automated feedback.

AICET has developed a programme known as Codaveri, which can help to provide feedback to students on their coding assignments. It does so by providing comments when it detects an error in their answers, as opposed to just giving them the solution. Teachers can then review the feedback provided by the AI and make adjustments or add additional prompts if needed. 

This tool is particularly helpful as there is a severe shortage of computer science teachers, Leong says. At the moment, AICET is working with MOE to run trials of the platform in schools that offer computing as a subject. 

While the programme today is able to provide preliminary inputs and feedback, Leong explains that it is still a work in progress. Teachers will still need to review the AI’s responses before it is ready for the students, but what AI can help with is providing some preliminary feedback to save teachers time and make grading more efficient.

Also read: APAC primed to unlock full potential of cloud with entry of AI-powered SaaS in Singapore