Doing away with teachers, well, something like that

Late last year, our education authorities welcomed Malaysia’s performance in the Programme for International Student Assessment (Pisa) 2018.

It assesses the proficiency of 15-year-olds in 79 high and middle income countries in reading, mathematics and science.

A beaming education director-general Amin Senin said we were not in the bottom third of all countries anymore but in the middle third.

That’s really no cause for celebration.

We’ve done slightly better than before, sure. But it certainly doesn’t merit the kind of pride Amin expressed nor the applause he received when he made the announcement.

We’re still very much below the Organisation for Economic Cooperation and Development (OECD) average, and are pathetically distant from chart toppers China and Singapore – with a more than 100-point gap separating us from them.

It reinforces the numerous complaints raised by parents about public schools – all of which have led to Malaysians becoming disillusioned with the national school system.

This is reflected in the fact that Malaysia now has the highest number of private schools per capita in all of Asia. We had 66 private schools in 2010. By 2017, it had ballooned to a whopping 126.

All this points to one thing – our national school system needs to be resuscitated. And it needs to happen sooner than later.

The biggest problem we have, as detailed in my previous column, is the inconsistency in and of our system. A few students receive a good education, but the vast majority don’t.

What we need to do is to be able to provide a top-class education for all our children. And a potent tool to do that is just becoming available to us: artificial intelligence.

The more evolutionary branch of this budding field is AI-assisted education. Instead of entirely upending the status quo, it aims to assist and enhance it.

One area which stands to benefit immensely from this technology is grading – a soul-sucking, sedentary activity that fills up a large portion of a teacher’s working day.

As is the case often these days, China is leading the pack on this front.

Almost a quarter of all schools in China now use an AI grading system. It is eerily powerful, being capable of doing things that only human graders could do just a few years ago.

In addition to being able to ascertain the logic of the text and judge an essay’s overall quality, it also provides recommendations on the writing style and theme. It’s also accurate, grading it as a human grader would 92% of the time.

This is, however, only the tip of the iceberg.

What will really move the ball are AI learning platforms that outperform human teachers such as ALEKS (Assessment and Learning in Knowledge Space) – a platform created by American education giant McGraw Hill that teaches courses all the way from kindergarten to university. It creates an array of “knowledge points” which it uses to track the performance and competence of students.

The big draw here is hyper personalisation. For thousands of years, students have sat in a classroom-like setting and learned from a teacher who educates dozens of students at once. This has been the case not because it’s the best way to teach or learn but because that’s what the logistics of the day would allow.

But not anymore.

Assessment and Learning in Knowledge Spaces, or ALEKS, allows students to go at their own pace and makes sure there are no knowledge gaps in their thinking. In addition to teaching the topic at hand, it also tests students on prior material to ensure memory retention.

Building on this foundation, Squirrel AI has devised the most ambitious of all AI in education platforms. The billion-dollar Chinese startup is looking to revolutionise education by creating the most powerful AI-led learning system to date.

It almost entirely does away with teachers with the help of its AI tutor, providing students with all the guidance and help necessary to ensure they excel at the subject of their choosing. It uses the same knowledge points system that ALEKS does, but puts it on hyperdrive.

For instance, Squirrel AI divides up middle school-level mathematics into 10,000 knowledge points, while ALEKS only divides into a 1,000. This disparity in granularity aims to ensure students don’t have any gaps in their understanding of the subject. And so far, the results have been encouraging.

A self-funded four day study involving 78 students showed that Squirrel AI’s system was better on average at improving the math grades of students compared with a teacher teaching a moderately-sized classroom.

Squirrel AI’s ambitious founder Derek Li goes so far as to say that “In three hours we understand students more than the three years spent by the best teachers”.

This might not be as outlandish as it sounds. One study found that spending a mere 34 hours on Duolingo, a popular language learning platform is equivalent to an entire university semester of language education.

One area of AI that will have an overwhelming impact on the industry is computer vision. Educational systems of the future will be fitted with cameras that possess machine smarts which will study and analyse a student’s eye movements, expressions and body language, using them to gauge their level of interest and proficiency.

By studying where and when students pay attention and how well their interest is held, these systems will be able to evolve rapidly, becoming more and more effective at teaching the subject at hand.

As exciting as these AI-propelled advancements are, the industry is still in its infancy and Malaysia can get a head start if we dive into the field now.

I see a future world where teachers will do less teaching and more facilitating while providing the much-needed secret ingredient that will guide our students – EQ or emotional quotient. They will provide the absolutely essential human touch that these AI systems lack.

Used in unison, I see AI and teachers holding the promise of a better education system, provided, of course, our politicians let the professionals handle it.

As a first step, the education authorities should select a core group of educators to be trained in AI-assisted learning, and prepare the groundwork for it.

Let’s hasten it, shall we.

This is the second part of an earlier column: Science’s solution to a broken school system?

The views expressed are those of the writer and do not necessarily reflect those of FMT.