Skip to main content

Get on Board with AI

Anant Agarwal discusses how and why educators need to embrace AI
Artificial intelligence concept art

That Artificial intelligence (AI) is rapidly changing how we work has profound implications for the future. A recent study, conducted by edX Founder Anant Agarwal and Workplace Intelligence, reveals how AI is already impacting the workforce. With the explosion of AI, 87 percent of executives are struggling to find talent for jobs.

In order to prepare students for the future, educators must also learn to incorporate AI in their classrooms, Agarwal says, and compares the adoption of AI in education other technological innovations like the internet. 

“While students should learn how to use AI to research stuff and … find different kinds of content, we need to help them understand how to use it in their day-to-day lives and at work,” he says, “much as we brought in search engines and the internet into the education process without fighting it." 

In this episode of the EdCast, Agarwal discusses why everyone needs to upskill in AI and how educators can begin learning more about AI in order to better address it in their classrooms.  


JILL ANDERSON: I'm Jill Anderson. This is the Harvard EdCast.

Anant Agarwal says everyone needs to rapidly upskill in the art and practice of artificial intelligence or AI. He's the founder of edX, and a leading thinker on technology and education.

There's been a lot of development in AI over the past six to eight months. He says we are struggling to figure out what it all means and what students should be learning, but it's crucial for our future. A recent survey he conducted showed 87% of senior executives are struggling to find talent to do jobs, and are desperate for those skilled in AI.

I wanted to talk to Anant about what ramping up AI education looks like, and how educators can gain skills in this area. There's reported estimates that almost half of high school students will use AI this year, so I asked him what educators can do.

JILL ANDERSON: From what I hear, high school students are especially active and engaging with AI. They're using it for their schoolwork, for their homework. I saw some data out there-- I don't know how legit it is-- that said almost 44% of high school students this year will use it for their homework. So one thing we know is young people need to be learning about how to engage with AI in the world, and especially how to be successful in the future. And you recently worked on a study that explored the workplace in AI and what folks are going to need going forward. So I want you to talk a little bit about that survey to help us figure out what educators need to be teaching about this.

Anant Agarwal

ANANT AGARWAL: You know, Jill, you talked about 40% of students using AI already. The genie is out of the bottle.


ANANT AGARWAL: Those teachers, or education institutions, or governments that try to prevent access to the technology, I think, are being very myopic. I think we need to understand that the technology is out there. It's going to be in all of work.  

We partnered with Workplace Intelligence to conduct a survey of C-suite executives and knowledge workers, 1,600 of them, and the results are pretty staggering. 77% of C-suite executives are saying that AI is disrupting their business strategy. To me, the scariest statistic that came out is that 80% of the same C-suite executives say that they're struggling to find talent with AI skills.


ANANT AGARWAL: Already. And so they already need the talent, and 87% are saying that they're struggling with finding talent even as their businesses are being completely disrupted by AI. So AI is going to be there in every which way, in every facet of life and work, going ahead. And so we need to be exposing our students to AI in terms of learning about AI, and also using AI tools immediately. I encourage students and workers and everybody-- much as we all have a Google search window open, I believe people should have a similar search window into a AI chatbot. Whether it's ChatGPT, or Bard, or Bing, we should all begin to start using and getting used to how to use AI in our day-to-day work, much as we use search engines.

And the analogy I make is that when search came out and the internet came out 20, 30 years ago, there was similar concern across schools and colleges—‘oh my God, what's going to happen to how we teach history now?’ People are not going to memorize dates. They can just Google search and find any piece of knowledge.   

Indeed, that did happen. But our education system learned to incorporate the internet into the research students had to do to write papers and so on. So we focus on critical thinking, not so much on regurgitation of knowledge. So the search engines democratized access to knowledge, or democratize access to facts. However, critical thinking was important.

I similarly think that AI is democratizing access to writing literacy. But critical thinking and these concepts are evergreen. And so while students should learn how to use AI to research stuff and how to find different kinds of content, we need to help them understand how to use it in their day-to-day lives and at work, much as we brought in search engines and the internet into the education process without fighting it.

JILL ANDERSON: That seems like an enormous challenge to get educators up to snuff on how to even be effective in teaching about AI.

ANANT AGARWAL: So there's really two facets of AI as it relates to educators. One is, how do I teach AI and machine learning as a concept? The second one is, how do I teach people how to use it?

The former requires courses in AI. You learn about machine learning and AI. And not everybody has to learn the foundations of how machine learning works. And people in tech and software have to rapidly upskill with AI because they are developing these tools and technologies, and so they have to learn about AI. And so for that, educators, universities, colleges, platforms like edX, have to rapidly create content that teach people about AI.

So as you might imagine, we've launched a number of programs. We just launched a boot camp in artificial intelligence. The boot camps can take people with very little knowledge in the field to enough knowledge that they can get an entry-level job. We've launched boot camps. We've launched the master's degree, as I said. So these programs will help people to develop AI tools and technologies.   

The other facet is, how do you use AI, whether in the classroom or at work? That does not require incredible skill. That requires a concept called "prompt engineering." Prompt engineering is an approach where you figure out how to ask questions of AI tools so that they can help you with what you want. AI tools cannot divine what you're thinking, so you have to ask, text, or other questions, or verbal questions, and AI can help you.

And so a rapidly burgeoning field in the application of AI is called prompt engineering. And as you might imagine, we have several courses on prompt engineering on edX. Prompt engineering-- let's say, for example, you are a manager. And you've been asked to give a presentation at an external conference, but you don't have the time. So you want help with writing a letter declining the invitation.

So you can use prompt engineering to ask ChatGPT, hey, can you draft me a letter declining an offer of this kind? You make the prompt, and you get a draft. And then you have to improve the draft. That is a way of using AI. And prompt engineering makes you good at using AI.

And that is not very hard. Even very basic courses that we have on edX in prompt engineering can help you do a good job of that. And frankly, even if you haven't taken any courses, you can just converse with AI like you're conversing with an expert on a topic. You can get by just doing that.

So as a teacher and educator, you can begin thinking about AI as an expert, an omniscient expert, that knows a lot about a lot of things, and to think about how you can incorporate that into your normal teaching. And the analogy I like to use is that when the internet came out, the internet knew about a lot of things.


ANANT AGARWAL: And teachers were able to incorporate that in their teaching. And even before that, when the calculator came out, the calculator could do quick little calculations. And teachers began to use the calculator in their teaching. Similarly, think of AI as an expert that can answer conversational questions and give you answers.

And so teachers can begin thinking about how they can use this expert, how students can use the expert. Teachers can use this expert to help them prepare lessons. On edX, for example, we launched a tool called edX AI summarization that allows teachers to very quickly create summaries of their lessons for the students. So there are many ways in which teachers can use it. And similarly, they can also give students assignments that assumes the presence of an expert that students can work with.

JILL ANDERSON: So it sounds like on the higher ed side of things, those institutions can be working to create courses, create professional development opportunities out there, for folks to expand their knowledge. And then it sounds like on the ground, maybe in K through 12, that this is something that's really easily accessible for people. And if you're sort of a tech phobe like me, you can really grasp this. It's not as hard as it might sound.

ANANT AGARWAL: It is not at all as hard as it might sound. And frankly, I think one of the big contributions of ChatGPT was making AI completely accessible to the common person. For anybody that's worried about it, I would simply say, use the internet analogy. Imagine a teacher's world before the internet, and then imagine a teacher's world after the internet. And how do teachers adapt to that? Similarly, there's before and after AI.

And so I would say that teachers can readily begin using it. And we're also working on a course for educators on how to use AI effectively in the classroom. And we hope to come out with that course in the not too distant future.

JILL ANDERSON: But I'm wondering if you've observed any ideal models of incorporating AI into classrooms that have really stood out to you.

ANANT AGARWAL: So these are, again, extremely early days. And I don't think there have been any key approaches in which teachers and professors have used AI.

In our own work, we have used AI in our courses in multiple ways. One is, as I mentioned earlier, we've launched a summarization mechanism. Our tool is called edX Xpert. X-P-E-R-T. And expert summarizations provide summaries to learners very quickly. And learners are saying that in many cases, they're learning more from the summary than from the video or a large body of text.

A second example that we've launched is the edX Xpert Tutor inside courses, where students can engage with the tutor and ask it questions. So let's say I'm stuck on a concept. Today, if you're in a classroom and you're stuck on a concept, you have to go find a teacher or a TA. And sometimes, when you're stuck on a concept at 3 o'clock at night, you know, heaven help you finding someone to answer or give you help.   

On edX, we do have a discussion forum where people can type in questions, and other students can help you. So online, there are ways in which you can help each other. But now we've launched edX Xpert Tutor, where learners can ask questions at any time, day or night, and get instant responses.

But there, we've done the prompt engineering such that the tutor does not answer a question. I mean, can you imagine if the learner simply takes a question from an exam or homework and asks AI, hey, give me the answer to that? That defeats the point of assessments. So there, AI then asks the student, OK, do you understand this concept? Why don't you go learn this concept? Here's a place where you can learn it. So there are ways in which teachers can use AI to help students even when they are not they are not around.

Another example that I think teachers can use is, before AI, let's say I have to write an article on something. I would create an article from scratch. Now I know that I have incredible difficulty, what you might call start-up trouble, with starting writing something. So I might create an outline, and then I kind of get stuck.   

A new approach to writing might be how people develop programs, where very rarely do people write programs from beginning to end. They iterate and make the program better and better. So one could imagine giving an assignment where — let's say in some field, you give an assignment. And you tell the students, OK, you can use ChatGPT's help, but then you have to personalize the essay.

So you can get ChatGPT to give you an early draft of something. It will have a lot of facts and features. But then you can take it, and then you can edit it and make it much more in your voice, make it accurate, and make the narrative be much more along how you want it to be. So that could be a new way of writing that I think is going to become very popular where, rather than creating something from scratch, you get a draft from AI, and then you iterate on it till you get it into a shape that you like.

JILL ANDERSON: You know, as we're sitting here talking, I keep thinking about your survey. The data — did you say 87% of executives are not finding people already that they need to do the work?

ANANT AGARWAL: Well, 87% of C-level executives are saying that they are struggling to find talent with AI skills. Simply put, everybody needs to rapidly upskill in the art and practice of AI. So for example, if you are a coder today, if you write software, you need to learn about AI very quickly and learn how to use AI in your work in software. You need to learn how to develop software using AI, and so on.

And it doesn't take another degree to do it. You can upskill with a short program in that topic. Let's say, for example, you are a journalist. You need to learn very quickly, how can I use ChatGPT and AI to become a better journalist? So it doesn't matter what field you're in. No matter what you are — let's say what you are is X. Everything is going to be X plus AI. So everybody has to upskill and figure out how to use AI in whatever they're doing.

And frankly, there are many courses already. Short courses. We launched a AI MicroBootCamp for coders where, if you have coding experience already, you can take a 6-week or 8-week MicroBootCamp that you can very quickly take and upskill enough that you can get a decent understanding of AI. Of machine learning, and deep learning, and so on.

JILL ANDERSON: What about leaders of school districts and colleges and universities? I'm wondering what their responsibility is to really get some movement happening in their districts and on their campuses in this?

ANANT AGARWAL: You know, there's a different level of AI that leaders have to learn about. They don't need to understand how to use AI in writing code, but they need to have a high-level, leadership level, C-suite level understanding of AI in terms of, what is AI? What can AI do for them in business? What will it disrupt? What will it not disrupt? What are its strengths and weaknesses? And probably most importantly, what are the ethics by which they deploy and use AI?   

So leaders need to understand that very quickly. Then again, there are executive level AI courses and programs that they can very quickly take and upskill themselves. So it doesn't matter who you are. Whether you're a coder, whether you're a writer, whether you're a psychologist, you're a CEO, or a CHRO, you need to get appropriately upskilled with the right kind of AI knowledge yesterday.

JILL ANDERSON: Right. Get on board, or you're going to get left behind-thinking?

ANANT AGARWAL: That is correct. We've been talking about the future of work and how half the jobs — the World Economic Forum's report that by 2025, half of today's jobs will be transformed or gone. Now, this was before the whole AI revolution happened a year ago. But the AI revolution — suddenly, even more jobs are at risk.

And so our society over the past many years has become used to an incredible pace of technological advancement, so that upskilling and reskilling have become part of our lingua franca. And so we knew we've had to upskill and reskill on a continuous basis. And now with AI — well, there's one more thing we have to upskill ourselves with now.

JILL ANDERSON: Almost seems like the good news, though, is that we do kind of have an idea of how this works now because we have been exposed to so much advancement in technology.

ANANT AGARWAL: That's right. I like to joke that in 2008, the word "data science" had not been coined. If you talked to somebody about data science in 2006, 2007, they wouldn't know what you're talking about. So data science as a field was coined in about 2008.

And I still remember at MIT, we were brainstorming. Should we be calling our effort in this area big data, or should we be calling it data analytics, or should we be calling it data science? And ultimately, data science became kind of the phrase that took over.

And everybody was looking to upskill with analytics. If you were doing marketing, you needed to learn about marketing analytics. Every field was being transformed because of data and the internet. So similarly, with AI, now it's one more thing. The past 10 years has shown us that we could not go and just get an undergraduate or graduate degree and rest on our laurels for the rest of our life. We have to continuously be upskilling and reskilling.

And corporations are understanding that as well, where they're coming to us and saying, hey, look, can you give us a catalog of courses that we can use for our employees? So we just launched the edX AI Academy, which is a catalog of courses that a university superintendent or college superintendent could buy for their school district teachers, or a CEO of a company or a head of human resources can buy for their employees.

So these are the kinds of packages that are in huge demand right now for upskilling and reskilling. Human resource departments, whether in school districts or in corporations, are completely aware that upskilling and reskilling continuously is a thing.

JILL ANDERSON: Any final words of wisdom for educators out there who are embarking on a new academic year, whether it's in K through 12 or in higher ed?  

ANANT AGARWAL: It's actually not hard. The simplest thing you can do is just go to a site like and type in "ChatGPT." And you'll get an intro to ChatGPT course. It's one or two hours. Just do the course. It's simple. It will give you a nodding acquaintance to ChatGPT. And then you can do a course on prompt engineering. And you can do both for free.

And it won't take you more than a couple of hours to do them. And immediately you'll begin seeing ways in which you can talk more intelligently about AI, you understand it better, and you can also help others use it better.

JILL ANDERSON: Thank you so much. This has been so helpful in gaining some understanding and almost making it seem less scary, I think, for a lot of folks.

ANANT AGARWAL: I totally agree, Jill. This is not scary. I think the fear has come from not knowing about the unknown. When calculators first came out, people were concerned. Oh my God, what's this thing? I remember my dad, he would complain, oh my God. He was a professor of pathology in a medical college. And he says, I don't know what this is, what is this thing, you have to punch these numbers, and so on. You know, just give me a paper and pencil. But once he overcame his fear, he was able to use calculators. And you know, of course, right? 

I think using AI is as easy as using a calculator. In fact, it's easier, because you can talk to it like a human being. You don't have to use a finger to hunt and peck. You can just talk to it.

JILL ANDERSON: Well, I'm certainly glad that the calculator was invented, so maybe in a little bit of time I can come around to using ChatGPT more.

ANANT AGARWAL: One piece of advice I'd give everybody is, much like you have a Google search window open on your screen at all times.


ANANT AGARWAL: Have a ChatGPT window open on your screen at all times, and just start playing with it. It's free.

JILL ANDERSON: Anant Agarwal is founder and CEO of edX, and a leading thinker on technology and education. I'm Jill Anderson. This is the Harvard EdCast, produced by the Harvard Graduate School of Education. Thanks for listening. 


An education podcast that keeps the focus simple: what makes a difference for learners, educators, parents, and communities

Related Articles