Welcome to Educated Voices, an interview series dedicated to K-12 educators and their work navigating the arrival of AI. I'm Emma Lewick. And I'm Bailey UAC. Our goal is to explore the question: how do we prepare children in a world that is fundamentally shifting? To answer that, we're looking outside the classroom to those leading, researching, and working on the edge of innovation. Our guest today is an AI pioneer who has quite literally taught the world how AI works. Dr. Andrew Ng is a founder of deep learning AI, a leader at AI Fund, and a co-founder of Coursera. He's also an adjunct professor at Stanford and was the founding lead of Google Brain. He is truly a luminary in the field. Andrew describes AI as the new electricity, a transformative force that will soon power every industry on the planet. But more than a technologist, Andrew is an educator at heart. He has mentored millions of learners. And today, he's here to help us understand what it means to be AI literate in a world where the rules are being rewritten. We are honored to have Andrew join us today to bridge the gap between computer science and the K-12 classroom.
Andrew, thank you so much for joining us today. We're really excited to learn with and from you, to share with our K-12 educators who are really trying to navigate AI for the students and for the possibility of tomorrow.
No, thanks for having me. I think K-12 educators are some of the most important, sometimes unsung heroes and heroines of society. So thrilled to be here.
Thank you so much. So to start us off, for a K-12 educator who is new to this subject, given your deep expertise in generative AI, where would you suggest they start?
One of the challenges I see in the whole K-12 system is it's under-resourced relative to a lot of other parts of society to figure out where we're going with AI. So I spend a lot of time talking to large businesses, to universities, and even the well-resourced businesses feel like they're behind and struggling to catch up. And so when I speak of leaders in the K-12 system with far less resources compared to these businesses, I see the struggle to figure out what's going on and keep with the times. Having said that, the good news is AI technology available on the internet is actually accessible relatively inexpensively. And I think that when teachers start to experiment with AI models, be it ChatGPT or Gemini, Claude, or others, and start to use them, it could already help with the day-to-day work of teaching. But then the other harder piece I think is really important is helping our students navigate, frankly, not the jobs of today but the jobs of the future. And this means helping them grow up into a future where they will be assisted by AI. Putting in place a computer science program to teach kids to code — I think those programs will be important.
Yes, absolutely. What is it about coding that a student gains or will take with them into the future?
One of the most important skills of the future is the ability to tell a computer exactly what you want — so, computer, do it for you. And for the foreseeable future, the language of computing is coding. So what I'm seeing already is that for a lot of job roles, people that know how to code — be a marketer or sales rep or HR professional or finance analyst — they could do more than someone in a similar job role but that does not know how to code. And I think even teachers — I was speaking through one of our teams, Kira Learning, which does a lot of work in K-12, speaking of one of the teachers that is a basketball coach and teacher, and because he had learned to code, he was writing data analysis to help his students visualize their basketball outcomes. So he was a better teacher because he could code. And so with a little bit of time to help people to transition, but I think in the future, maybe nearer than people think, when you code, I think you could be a better teacher as well. And the trick is, of course, you don't have to do all the coding yourself. You get AI to code for you. So one quick example: one evening my daughter, she wanted to practice multiplication tables.
And so, you know, I did the latter. It took me an hour. Then my daughters spend time practicing. And so even as a parent, I'm able to better engage, you know, my own kids because when there's a specific task, I get the computer to do it. And I think it'll be very true for a lot of teachers as well.
I really appreciate you sharing that. And as an elementary school math teacher, I had come across that same thing and being able to use that coding to create a multiplication — funny, it was the same thing as well for the students. And then I actually had them give me feedback so that we could make that better and share it back with them. So just one of those ways that they got more agency out of it as well because they were developing the tool to help their own brain learn and grasp these concepts. Can I ask you, as a parent raising your own children, what kinds of educational experiences or life experiences are you really planning for your own children or that you really want to make sure your own children have?
I honestly wish I knew. I think some things are becoming clearer. So when I look at K-12, also through college, children and even ourselves, we will live in a world where AI can help us. So this means that information search and discovery, that will be AI models. And then it turns out, maybe looking to an earlier era of technology evolution with the rise of the internet — Google won the horizontal information discovery layer. Then there were verticals like travel, retail, transportation, and so on with different businesses formed out. I think the new value landscape is still emerging for this wave of AI, but the horizontal information discovery — ChatGPT, Gemini, and so on — that's becoming clear. The one vertical that's becoming clear is coding. And I think everyone, you know, getting AI to code for you would be more productive. The other verticals — what's the future of marketing, what's the future of sales — those are not yet clear. But I think that horizontal information discovery, question answering, companion kind of assistance, as well as writing code, that's becoming clearer. But the other disciplines are still coming into focus.
So having K-12 educators really think about the skill of coding and breaking that down into kind of the step-by-steps that students at each grade and level could use and practice as a way of thinking for the future and preparing for the future.
Yeah. And you know there's actually one funny result from the AI world, which is it turns out that if you train the AI model to code, it actually gets better at other tasks as well.
And so, you know, we can hypothesize why, but coding involves so much logical thinking processes. I suspect that when AI learns to code, it becomes more logical in instructions and other things.
Part of me thinks it may be true for people as well.
I think positive byproducts. We would agree. I think that the more we engage with children, the more we teach them to think, we see gains in a wide array of places. So and I know that there were people including senior leaders who were advising others not to learn to code on the grounds AI will automate it. We'll look back on that as some of the worst career advice given. And what I'm seeing at, say, the college applications already — not that life is about college, but this one important thing — I see a lot of high school students that learn to code and they can just create more. And so to me, I often think about how do we empower children to create and to give them the foundational building blocks. You need the hard skills and the soft skills. And I'll just share one more thing.
I think for strange reasons, there's a lot of hype and fear about AI. Maybe two things. I think around 2023, ironically, right as AI was taking over the United States, some companies ran PR campaigns to say AI is dangerous. It can even lead to human extinction, which I think is completely ridiculous. One really unfortunate impact of that is my friends and I, we've spoken to K-12 students that have said things like, 'I heard about AI but I don't want to work on it because it may lead to human extinction. I don't want to contribute to that.' Yeah. And so I see that hype as incredibly damaging because it is dissuading young students from entering what could be one of the best, most transformative moves for their lives and their careers.
So, encouraging our students to embrace this wonderful technology and not buy into these very bizarre fear narratives — I think it would be important for us as educators to do as well. Do you have any favorite resources that you would recommend to teachers, to families, to students to rely on to inform themselves about AI and kind of steer away from the hype to help them do their own research, right? Because we want them to take that and understand this narrative themselves.
Yeah, that's a really good question. I wish I had better resources. So at DeepLearning.AI, we're committed to empowering everyone to build with AI, and we have a good record of being quite factual, I think. And so much of the information ecosystem is so noisy and so polluted by marketing hype. I find that if you have just a little bit of technical understanding of how it works, it makes it easier to sort through some of the noise. It's worth it, just like today, everyone's kind of got to learn to use email. I can't imagine — I wouldn't want my kids to grow up and not learn how to do web search or use email. And I think there's that technical understanding that is needed.
You know, change is hard and all that, but people will need to learn this. Yeah, I would say, you know, we started with your AI for Everyone course.
Yeah. And having that little bit of technical information really opened, I think, both of our eyes to the possibilities of what could be. And being able to see past a lot of the hype to reimagine education and to reimagine what the educator's experience could be and also the student's experience could be, and the structure itself, right? Generative AI has given us the opportunity to take a step back and see what do we need for our students. What do we hold true of child development, of learning science, and then what do we need of our future and where are we heading. And so we feel this pivotal moment and we're really appreciative of the work that you do and that we're excited to be here at the table and to be here for and with our K-12 educators. So you encourage people to build with AI to unleash innovation and a lot of what you're talking about, being able to have that coding and the skill and the building and the creativity. So what are some of the most striking educational applications of AI that you have come across?
So there are a lot of them. I don't know that there's one that completely stands head and shoulders above others. I feel like AI-oriented chatbots have been very helpful. So for example, Coursera Coach is a very effective chatbot that helps learners get to the answer by themselves rather than gives them the answer. So that pedagogical learning journey I find very helpful. And then I think that Coursera Course Builder also helps businesses or teams assemble new curricula. At Kira Learning, which is a company that I've been helping — I actually know the co-founders really well — it has been using AI to help generate some content to make it more efficient for the teacher. I think the flipped classroom model has been discussed for a long time, and I love the concept. I even do it myself sometimes in my university teaching. It's never really taken off. And I suspect that one of the reasons why the flipped classroom kind of worked, kind of didn't, was because of the lack of personalization. But with AI, we can now actually write software — Kira Learning has this, it drives hyperpersonalization — where the AI can look at a student submission and help a teacher pinpoint, you know, this student got confused about this one line. And in fact, if a teacher is busy juggling too many things, it can auto-suggest to the teacher, 'Hey, if you say these words to the student, you can help them get unstuck.' And if you can do it or not, right? But you're getting really positive feedback from teachers on how it empowers them to very quickly target exactly what they need to do to unstick a certain student.
I think that's wonderful and really looking in this space where in education we know these best and promising practices and with AI we're able to actually implement them, right? Giving that individualized learning, the one-on-one, the coaching, and even coaching the teacher in the moment when they're working with the student. I think that's just wonderful. I really appreciate you sharing that.
Another important role is, you know, the content delivery. But it turns out that when AI or digital media can take over more of the content delivery, the teacher can spend much more time in social-emotional support, and that is incredibly valuable. We're really excited for teachers to get the opportunity to really become human development experts and learning experts, right? And to be able to kind of make a lot of the invisible work that they do visible because they can give so much over to AI to support assessment, to support planning, to give them that time back to really be able to do some incredible teaching.
Yeah. And I would say it turns out, despite hype about AGI, artificial general intelligence or whatever, for the foreseeable future there's a ton of stuff that teachers do that I don't see AI able to do. And so giving teachers that time to create a more child-centered experience — and I look at technology, yes, social media feeds are very person-centered, but really it's such a shallow understanding of the present. It's such a weird thing, but that human, highly empathetic teacher with the time — that's the tough thing — time away from the algorithm.
Yeah. Can I ask a question? Many educators, K-12 educators, you know, say that they feel inundated right now with frameworks and tools and research and the news. Do you have any wisdom or advice to share?
One of the things that I did at DeepLearning.AI and Coursera was have a course, Generative AI for Everyone, that tries to convey to non-technical individuals a slightly technical, application-business-level understanding of AI. One of the difficult things about AI is it's fundamentally a very technical subject. And so nuances about what exactly the technology can and cannot do is not some esoteric piece of technical knowledge — it's a piece of knowledge that drives very concrete decisions by teachers and others to decide, should I do this or should I do that. So I find that to navigate AI successfully, not everyone needs a PhD in computer science, but having just a little bit of that technical understanding lets you very quickly decide: is this hype, is this real, is this working now, should I wait two years before I deal with this? So I hope that if teachers take Generative AI for Everyone, that conveys that type of understanding, and then it makes it much easier to sort through the noise.
