Atlassian Team ’25 – The founder of the Khan Academy says that AI will personalize learning for each student

The potential IA impact on society is breathtaking, but perhaps its influence on education and learning will be the most dramatic and perhaps the most useful.

In the latest speech of the Atlassian Team '25 in Anaheim, California, last week, Ben Gomes, SVP Learning and Education at Google, and Sal Khan, founder and CEO of the non -profit organization of education Khan Academy, discussed the impact of AI on education and learning. They both believe that AI will play an enormous role in scaling up learning, promoting curiosity, encouragement of critical thinking and improving educational results.

Gomes, who joined Google in 1999, stresses that, before the arrival of search engines, obtaining information meant to go to the library, and it was often difficult to find the information you wanted. He continues:

Google has decided to try to solve this problem. The first time I really had an idea of ​​how it changed the world was when I was on a bus trip between Bombay and Pune around 2000, and several people on the bus used Google. Now it seems completely normal, but at the time, there was nothing on which I worked who had this kind of reach. There was this enormous transformation.

Sal Khan fell into accident education 20 years ago after killing his cousins. Before he knows, he worked with more than 15 children and thought that the use of technology would help him follow. He says:

I have seen a model that you may be able to use the technology, in one way or another, to help people learn. So I started writing software, it would give them a practice and comments. But I continued to explain the same thing.

A friend suggested that instead of repeating himself, he could make videos and put them on YouTube. He started making rough videos explaining things he had repeated to his cousins, and sent them an email with all the questions they had. The comments he obtained were that they preferred tutoring videos in person. He explains:

They loved nature on demand. There was no shame if you were an algebra student and you had to review the division decimals, or watching a video at 11 p.m. and repeating it.

Interactive learning

Khan discovered that there was still a problem – if a student watched the video and still did not understand, he would be stuck. The generative AI will be the solution to denigrant students, according to Khan. He explains:

We are looking for projects where General AI is now at a point where he can ask you a few questions and try to focus on what you need. This can make it more interactive.

The two speakers agree that the release of teachers will also obtain a huge boost from the AI. Khan says:

Students can ask questions that help teachers at course plans, progression reports and possibly papers. I think this is the first phase, and the next will be to make AI more proactive …

Helping teachers in planning, delivery and evaluation can link things, so that they can create course plans, quizs and content, deliver it to the student, then get real ideas.

Khan believes that some of the education problems are due to narrowness of the evaluation. He says that when the system focuses on multiple choice, this means that assessments cannot examine critical thinking or problem solving – but AI will be able to make these types of evaluations in the future.

The Khan Academy is launching an AI simulation product in this area next year. For example, if a student had trouble understanding the causes of civil war, a teacher would be able to design an assessment with an AA simulation of Abraham Lincoln.

Socratic learning, where students learn from discussion and questioning, is a perfect match for AI, according to Gomes and Khan. Gomes says:

I think what really sounds with me is this notion that the best teachers are able to do it – and many other teachers would like – but it takes a long time. Obtaining this kind of experience for teachers and students would really change things.

The two speakers agree that curiosity and motivation are the foundations of learning, but that they must be nourished, and AI will help to do so. Gomes says:

Curiosity is our instinct to learn. You were born with her in a deep way. But you need advice, because curiosity dies if you cannot get any answers to your questions.

I felt that research had helped you go to the next step, just like Youtube. But now I can ask these questions at AI, and AI can also start suggesting questions. It is a proactive experience that helps to advance this curiosity. So you always learn but it seems effortlessly. It cannot be effortless, but it should seem effortless.

Personalized education

The more interactive it becomes, the more personalized education can be, according to Khan, which emphasizes that all the education has been personalized at the origin. He explains:

Very few people had an education, but those who did it, obtained a personalized it. Alexander the Great's Tutor was Aristotle. And you can imagine that when the young Alexander was ready to accelerate, Aristotle gave him something more difficult. So it has always been the gold stallion, but it was expensive.

He thinks that mass public education, although extremely positive, has noticed compromise, such as gathering groups of students by age and trying to progress them together. We hold this form of education for acquired, but it was much more personal. Khan continues:

Some children learn quickly, but some have gaps. These gaps accumulate and at one point, they hit a wall. If you are middle class, the way you got around is hiring a tutor. What is exciting now is that there are so many more opportunities for evolution of approximate personalization via AI, and as it improves, it will diagnose where these shortcomings are.

He thinks that over the next two years, AI will lead to a very proactive approach to education. He adds:

I think this is where personalization goes, and of course, for a teacher with 30 children in class, this level of personalization is really precious, because the teacher could never do all this by themselves.

Gomes agrees, adding that personalization has many levels – to help teachers help students working at different rhythms, to give them motivation to succeed in a material that they may not initially appreciate. He says:

I will try to understand the tools that can help the teacher to that. But if you go beyond that, where the individual fails because he is not motivated and use AI to give them a goal, I think we will use it to understand the motivation for each student and to unlock it.

Discover Diginomica's Content hub of the dedicated team '25 here.