With the rise of Artificial Intelligence (AI) and Natural Language Processing (NLP) in Southeast Asia, there’s plenty of interest in the use of AI and its potential to transform schools. Many people are framing now as a time for “the future”, but how does it benefit the education scene? Should these technologies be used to assist and optimise the students’ education process? Is there even a future for such tools and technologies in the region? The short answer- Yes.
The increasingly rapid rates of AI adoption in the different industries within Southeast Asia are projected to increase the region’s GDP, adding $1 trillion, by 2030. The incorporation of digital tools and technologies in and out of the classroom have seen schools experience between an 11% to 28% rise in various benchmarks, namely efficiency, funding, competitiveness, innovation and student engagement.
Additionally, the funding for AI in the education sector is set to almost quadruple, higher than most industry sectors within the Asia-Pacific region while the global market for AI in education is forecasted to be USD$3.68 billion by 2023.
NLP, which gives computers the ability to understand text and spoken input similar to humans, is expected to grow at a higher compound annual growth rate compared to other AI sub-technologies like machine learning and deep learning by 2023.
We speak to edtech founders about the industry in 2021
With plenty of applications, adopting these technologies into the education process will greatly expedite students’ knowledge acquisition, complementing existing classroom teaching methods and decreasing teachers’ workloads.
Personalisation of learning and knowledge acquisition
In order to provide a personalised learning experience for students, AI and NLP identify possible gaps in the students’ knowledge, using diagnostic testing or frequent assessments. It can then develop a personalised curriculum based on the student’s specific needs.
AI can also be used to give students reading materials that are suited to their levels. This can be done with the algorithms being able to compute the reading difficulties of open text data from the internet before recommending the material to the student. This application could be especially effective when students have already interacted with the AI and the AI has had enough data points to ascertain the student’s individual level before comparing it to the difficulty level of the materials.
Students in SEA have different language abilities and they all learn at different paces. For students that struggle with the language of instruction, the use of NLP in EdTech platforms empowers them to practice their language skills independently on the platform, without fear of taking up too much attention in the classroom. Additionally, because of NLP, the AI will be able to understand their words and interact with the student accordingly. At Noodle Factory, student interaction with the conversational AI interface helps to clarify their doubts and questions. Any weaknesses or gaps in learning can be identified and highlighted to the teacher within the analytics dashboard for their students.
Throughout the whole process, progress can be monitored by teachers, making them more well-equipped in identifying students who may need more support in their educational journey. Educators are then able to create personalised and targeted learning materials to help bridge the gap.
NLP may also be used as a gauge to measure the emotional state of the students in areas such as their self-esteem. As a result, this information can be applied to detect students at risk of mental health issues for early intervention.
Decreasing workload through a knowledge base
Content created by teachers, such as learning materials, can be uploaded to a knowledge base which can be a key resource for students. As students can readily access the learning content anytime through the knowledge base, it better facilitates their learning at home and can drastically reduce students relying on their teacher for help. This also helps to free up the educator’s time and can devote their time to other tasks or help struggling students.
The knowledge base can also be utilised by the AI, using NLP, to create assessments for students with ease. Teaching materials, as well as any documents provided by teachers and educators, will be scanned by the engine, to generate both short-answer questions and as well as multiple-choice quizzes based on the key concepts covered. Students interact with the tutoring knowledge base and assessment questions using a conversational AI interface, which again uses NLP to understand what students are asking, and then provide the tutoring and tests from the knowledge base and question bank. This lessens the burden on the teachers to create new assignments and also acts as tutoring and support for students outside the classroom.
Automation of marking and feedback
Research by McKinsey & Co. found that teachers can save 20% to 40% of their time by automating some of their tasks with AI. One of the key areas that can be automated (or at least made easier and quicker) is the marking of assignments. I’ve seen educators save as much as 400 hours in a year just by having an AI assistant help in the grading of her student’s assessments.
These programs can compare the answers between students’ submitted answers and the model answer uploaded by the teachers, deducing the reasoning and logic behind a student’s answer, instead of just merely looking for keywords. This is very effective for the open-ended type of short-answer questions and it helps teachers save time.
Additionally, AI has reached a point where it’s capable of providing human-like feedback on student essays with platforms like OpenAI. While not meant to replace a teacher reading and grading one’s essays, such software can give immediate feedback to the student. However, training such systems would definitely require some resourcefulness with the data collection but if done right, it could lead to ingenious applications that could cut down teacher workload.
Drawbacks of NLP
Unfortunately, most of the datasets for NLP and AI are in English, with not many datasets catering for the other languages within Southeast Asia. This hinders the potential growth and training of the AI systems since algorithms need large collections of data to train with and become more optimised over time and as more data gets inputted into the system. Moreover, the AI platform has to adapt across different geographical locations with diverse slangs and colloquialisms, and the interplay of cultural influences. Subtle nuances in languages could end up being overlooked.
When it comes to language and syntax differences, essay scoring websites had the tendency to award higher marks based on the complexity of words used and essay length and often deducted marks for poor grammar, style and organisation unlike when the same essays were graded by teachers who awarded marks based on the quality of the work, awarding higher marks for the latter criteria. As a consequence, students may get unfairly graded for their assessments if grading relies solely on automated technologies which are not able to identify nuances and quality of writing. Therefore, teachers need to check the grading system to ensure that it is fair and transparent to all.
Overall, checks and balances are needed since AI technology is relatively new, and AI should not take the lead in classroom lessons, but act as an assistant for teachers and support for students. Within Southeast Asia’s education space, NLP and AI definitely show a lot of promise for collaboration with educators, helping them to greatly improve the quality of education, complement existing teaching methods and relieve some of the pressure on teachers drowning in administrative work.
This was contributed by Yvonne Soh, co-founder of Noodle Factory
About the author
Yvonne is the CEO and Co-Founder of Noodle Factory, an AI-powered platform built for educators. She has over 22 years of experience in the technology sector, and spent many years in adult education, where she worked closely with different companies to incorporate technology to improve learning outcomes. Early in her career, she worked at multi-national corporations, including Dell and F5 Networks, where she was responsible for product management, development and systems consulting. She also spent a number of years in NCS, a large Singapore systems integrator, and was involved in the initial implementation of the Singapore government online infrastructure.
She is a self-taught technologist who loves to create awesome user experiences using technology and innovation. She is a lifelong learner. In her free time, she trains her dogs in tricks and agility, practices Brazilian Jiujitsu, pole dancing, and yoga.