Brave New Words: How AI Will Revolutionise Education (and Why That’s a Good Thing)
By Salman Khan
Viking
ISBN: 978-0593656952
272pp.

Artificial Intelligence (AI)-based digital platforms can provide access to quality education using personalised and interactive tutoring, and this can be accessible to a very large number of children and young people, at a fraction of the cost that would need to be incurred if we want to do it through the normal schooling model of 30-1 student-teacher numbers.

This is the main message of the book Brave New Words: How AI Will Revolutionise Education (and Why That’s a Good Thing) by Salman Khan. Revolutionising education is the main aim that has been driving Salman Khan as well. It is understandable why he is excited about the possibilities that AI opens up in education.

Salman Khan — of Khan Academy fame — is, of course, no stranger to education and/or online teaching and learning. Khan Academy is one of the largest platforms of its kind in the world and has been instrumental in providing access to quality educational content to millions of people, and Khan Academy videos have had billions of views.

Quality education not only requires good content that is conveyed to the student in interesting ways, it requires constant and interactive feedback for the student as well. The students also need to be encouraged, and inspired, and they need a lot of coaching too. Traditionally this was and has been provided by in-class teachers.

Khan Academy’s Salman Khan pens a book about the potential of Artificial Intelligence to make quality, personalised education accessible to more children at lower costs than traditional interventions

Khan Academy, before AI, was able to ensure the quality of the content and the way it was taught (pedagogy issues) but, Salman believes, Khan Academy was not able to provide a lot of the coaching elements. The interaction between the student and the material provider could not be very active, live, or responsive to the needs of the individual.

Feedback was provided in good time to the student and information regarding progress was conveyed to the parents and teachers, but it was not possible to have a very responsive and individualised experience of learning for students via the internet.

AI, Khan believes, has given a way for addressing the issue. AI, trained on the right education material, can provide immediate and effective feedback. Students would be able to have meaningful conversations with AI-trained respondents and this would be tailored to the needs of the individual. It will be like having a good teacher and coach with you.

This is the potential that Salman Khan sees in AI for education. He acknowledges that we might not be at this point yet. AI tools are still developing and getting trained. But the potential, for him, is definitely there.

Khan also talks about Khan Academy’s AI-assisted programme, Khanmigo, as a step in this direction. According to Khan, Khanmigo acts as a knowledgeable, interactive and supportive coach for students and it is at the individual level.

It is already able to provide fairly sophisticated interactions/conversations to students. Mr Khan gives quite a few examples of these, from various disciplines, in the book. But, more importantly, Mr Khan believes that the potential to develop Khanmigo as AI tools become more sophisticated, is what will make it even more impressive.

He mentions that they are involved in a number of studies that are looking into the impact Khanmigo is having and will be having on student learning and other outcomes. The results of these studies are going to be very important.

Quality education not only requires good content that is conveyed to the student in interesting ways, it requires constant and interactive feedback for the student as well. The students also need to be encouraged, and inspired, and they need a lot of coaching too. Traditionally this was and has been provided by in-class teachers.

These possibilities could be extremely important for countries such as Pakistan, which are struggling to provide quality education to all children. Pakistan has some 26 million five to 16 year olds who are out of schools. Most of the children who are in schools are getting poor quality of education. If we had the possibility of providing access to high quality coaching at reasonable cost to the millions of young people we have, this could be a game changer for Pakistan. But is this possible? And is this practically feasible?

Our internet infrastructure is poor. This will need significant improvements if ‘live’ interaction with an AI tutor is to take place reasonably. This cannot be done offline. This means we will need bandwidth as well as internet stability.

For children who are already in schools, it might be more feasible to do this intervention. A lot of schools have internet connections. Some have computers and devices as well. If individual devices are needed, this will be a problem. Many low-fee and public sector schools will find it hard to arrange for a device for each student. So, some device-sharing arrangements will have to be done. Or someone will need to subsidise device and maybe even subscription charges.

For out-of-school children, the issues are even more significant. A lot of out-of-school children are not coming to schools due to household poverty. Many out-of-school children are working as child labourers as well. Will they have the ability to afford devices and subscription? Will they have the motivation to persist? The state will definitely need to subsidise partially or fully.

Salman Khan does make the point that AI-assisted learning might still be cheaper than having a knowledgeable, motivated and capable teacher in a class setting. This might well be true. Even those who can afford to pay some level of tuition fee in Pakistan are getting poor quality education. So, the cost comparison might hold.

It is interesting that Khan iterates, numerous times, that AI cannot displace teachers. If a teacher is knowledgeable, knows how to teach and motivate, sure, AI might be a poor alternative. But if the teacher is not all that, and the AI alternative is better, will Salman Khan’s argument still hold? This seems unlikely.

So, AI will, at the very least, change the role of the teacher in class and, in some cases — though this is more likely in places where ‘real’ teachers are not as trained and/or motivated — might replace teachers as well. If Khan is right about the potential of AI-assisted educational programmes, he is wrong about the impact this will have on teachers and teaching. Irrespective, it might be years before this impact comes through: education is not an area where things change fast.

This is an important book. It points out the potential AI has in changing education and, possibly, improving it. Khan gives plenty of examples from different subjects and at different levels. The message is simple: AI-assisted education programmes can deliver personalised coaching for high quality interventions, and it is likely to cost less than in-person teaching.

It definitely has the potential to disrupt the status quo. For countries struggling to provide education to all, this could open new avenues.

The reviewer is a senior research fellow at the Institute of Development and Economic Alternatives, and an associate professor of economics at Lums

Published in Dawn, Books & Authors, August 11th, 2024