IEFG BIG Series: The Promise of AI in Underserved Education Systems

Welcome to the IEFG Brains in Gear series. In today's episode, we take a closer look at AI and its impact on classrooms, especially in low and middle-income geographies.

The conversation will be amongst the following speakers:

• Ana C Rold is an award-winning media executive and entrepreneur. She is the Founder of Diplomatic Courier, the flagship media network and magazine for top diplomats in Washington, DC, New York, Brussels, Geneva, and key capitals around the world. In 2012, Rold founded the global futuristic think tank/do tank, World in 2050. 
• Grace Cadiz  is the Partnerships (Philanthropy) Manager- Global Education at the Atlassian Foundation known for its commitment to improving education and supporting tech innovations.
• Asyia Kazmi is the Global Education Policy Lead at the Bill & Melinda Gates Foundation, with a focus on effective instructional practices, education advocacy and edtech. a global leader in philanthropy with a strong focus on enhancing educational outcomes worldwide.  
• Pat Yongpradit is the Chief Academic Officer for Code.org, a non-profit dedicated to promoting computer science education, and lead of TeachAI, a global initiative to guide education leaders in rethinking education in an age of AI. 
• Nissi Madu is the Managing Partner from Co-Creation Hub, an expert in leveraging technology to solve social problems and improve educational access across over 20 programs in 7 African countries.

Here are some resources if you would like to learn more about AI in Education:

• Teach AI's policy resources and toolkit 
• NORRAG AI and digital inequities summit
•  What will the future of education look like in a world with generative AI? | by MIT Open Learning | MIT Open Learning | Dec, 2023 | Medium
• Schmidt Futures-A Look at the Future of Learning Tools with Kumar Garg - Trending in Education
• Chatbot tutors have not arrived and we should wonder if they ever will
• AI in education: where we are and what happens next 
• AI as an arrival technology
• EGRA AI(Early Grade Reading Assessment- AI)
• AI

Subscribe to the podcast so you never miss an episode! And don't forget to rate and recommend this podcast to your colleagues.

You can follow the IEFG on LinkedIn here. https://www.linkedin.com/company/international-education-funders-group-iefg

Transcript

Pat: Will AI craft a new classroom pedagogy? No, AI won't do that. Teachers will do that and AI will help a little bit. 

Asyia: The other question is, is there an option to do nothing? I don't think there is. 

Nissi: How do we ensure that In terms of the data that's being used to train this, that it actually mirrors our unique context.

Grace: I see the Socratic method really being powered by AI, where students are encouraged to think critically. 

Ana: About half the world still doesn't have access to the internet. So we're over here iterating on rocket ships, but If we don't do the basics, like give them access, can we even talk about how revolutionary this technology is going to be?

Anjali: Welcome to the IEFG Brains and Gear series. In today's episode, we take a closer look at AI and its impact on classrooms, especially in the low and middle income geographies. The conversation will be led by Ana Rold from the Diplomatic Courier and two of our funding partners. Atlassian Foundation and Gates Foundation with views from two practitioners Pat Yongpradit from Code.org and Nisi Madu from Co-Creation Hub. 

Ana: Hi, I'm Ana from Diplomatic Courier and World in 2050, and today on the IEFG big series we will be talking about the technology that it feels like everyone is talking about artificial intelligence. Will it be the big disruption to the classroom model or education?

Can it help teachers with huge class sizes and wide variability in student competence? Can it help students to learn at their own pace and level? And will AI change the job market of the future? And how does that affect what children need to learn today today will explore how we think a I will shift the education sector and the questions philosophic funders are asking and should be asking we have with us to education funders and to practitioners.

Grace: Hi, I'm Grace from the Atlassian Foundation. Being a global education funder, I receive invitations to attend various events, read articles, and listen to new perspectives and ideas about how AI can play a significant role in education. I've come across some bold claims about the potential of AI, and I'm excited to explore this further today.

Asyia: Hi, I'm Aasiya Kazmi at the Bill and Melinda Gates Foundation. My background is 20, 30 years ago I was working with using computing, um, personalized adaptive solutions to improve learning outcomes for mathematics, as well as a teacher coach supporting teachers to use technology. So this is not new. But what I'm coming to this conversation with, as a funder, we ask ourselves, where will our investments have the most impact?

How do we know this is the place to make an investment? How do we know it works? And I'm hoping that this conversation, we're going to get in deeper into these issues. And I have two colleagues, buddies with me on that. The first is Pat Youngcredit, he's the chief academic officer from Code. org and he leads Teach AI, an organization that is working with multiple organizations internationally to lead the effort about teaching about AI and teaching with AI.

And Nissi Madu, a managing partner from Co Creation Hub. Um, which is a panel of time innovation enabler working across different sectors. Um, but especially the role on education today. Thank you so much.

Grace: Thanks for being with us, Pat. I'm super excited to dig into your work at Atlassian Foundation. We're big fans of both Code. org and the Teach AI initiative. Recently, Code. org launched the Teach AI initiative, which involves bringing education leaders and tech experts together to ensure responsible, safe implementation of AI in schools.

I was wondering if you could share your perspective on how AI can be explicitly implemented into education. 

Pat: I mean, that's a very, very big question. I think where I'll start is right at the beginning. What can education systems do to prepare their students and their staff For an enabled future and most people listen to this are probably new to a and so i'll say don't be afraid the best way to start might be just to listen rather than do anything and just listen to your stakeholder community your parents your teachers your students.

And then from there, from the listening session, start to develop a task force that involves all of those stakeholders to kind of define the strategy for the education system. And from there, move to light instructional guidance. to help dispel misconceptions about AI, to make sure that we addressed any growing inequities around the use and access to AI tools, and then also ensuring safe and responsible use of AI in the classroom.

Asyia: We're talking about such a wide variation within countries and across countries about knowledge about AI, tools that are available for AI. even access to power and connectivity. And so some of the things that we're thinking about are, how do you put a framework around this? We've come up with three categorization of student facing tools.

There's models coming of student tutoring. We just had some impact evaluation to show that's had impact or like products like MindSpark, OneBillion, Chimporn, all fit into that category. And then we have the teacher facing tools and the teacher facing tools, I think is perhaps where we have less evidence, but where I'm quite excited about what AI can do to support more effective teacher professional development.

improve support teachers to gain contact knowledge and pedagogical contact knowledge. And then finally, like system facing tools, so there's obviously this translation model where high quality content can be translated into less used languages so that students and teachers have access to this. Marley Robots is doing something on this, or Robots Marley is doing something on this.

But also, how do we speed up things and make teachers lives easier? So, Wadwani Organisation in India is doing one million assessments per month using AI. EGRA AI in South Africa is doing things like shortening the time it takes teachers to assess. So, there's use cases that are emerging that should solve some of the problems that we have in, in the global education context, that I think will be really good. Nissi , I know you're closer to this, for you to elucidate more on. 

Nissi: Thanks, Asyia, for, you know, providing that context on learner facing, teacher facing, and system facing tools, and just digging a bit deeper into the teacher facing bit.

We're working very closely with a lot of Government and national, subnational levels. And we're beginning to see, like you mentioned, that there's an opportunity for not just technology, but AI to play a very important role in that space. And essentially saying at the point where teachers need to deliver their lessons, can these tools serve as a productive assistant to just make it a lot easier for them?

And this productive assistant is not just. focus on pedagogical or subject matter knowledge, or even to very simple questions. Like I have 50 students in my class, for instance, and I want to teach this particular activity or what's the best format to group this activity in for me to be able to see the most value in it.

And so there's also the question around with the devices. Do we have the teachers use the devices that they have, their personal devices? Does the government provide that? Do we have a sort of community center where the teachers can come and access these tools and work together to find opportunities and ways that they can use that and then go back into the classroom and practice what it is that, you know, these tools have recommended for them to do.

And so those are use cases on the ground in terms of how we're thinking about, you know, applying. this and accelerating essentially what it is we want to achieve in terms of the learning outcomes and seeing a lot more children achieving, you know, literacy and numeracy competency levels at the required age and at present.

Grace: It really sounds like AI can, essentially what you're saying is take a Good learner into a great one through the feedback and adaptive learning. And then on the teacher side, it sounds like AI is really enabling them to allocate more class time to maybe what we call like HI, more human interactions rather than the administrative tasks like lesson planning, grading, et cetera.

Pat, I'm interested to see what you think is the most exciting potential of AI. So not just the use cases that you're seeing right now, but what do we have in the future in the classroom or in the global classroom? 

Pat: Sure. So the first real usage, one of my favorite usages of AI for teachers is just simulation.

There are often instances where you want to try something out, but you don't want to try it out on a human. You just want to kind of play it out. And even just the ability to have a metacognitive experience where you're trying something out, reflecting on it, but it has no consequence in real life.

That's a very useful thing. That's pretty cool. Now, as for my dream, I was a computer science teacher. I hope that an AI already can create code. pretty well, obviously not perfectly, but I hope that one day there will be AI that anyone can speak to and it can turn out anything. Imagine a generative AI that could create any type of application.

And so whether Nissi has a good idea or Grace or Asyia or whoever has a great idea, they can turn that into a functioning product. So I hope that will come about because what that will do is it will turn the entire world, every single person, into a creator of potentially world changing technologies. So I'm excited about what kids will create, what adults will create, what people will create for the communities.

Asyia: I want to jump into what Pat said about simulation. I think people really underestimate the demands we make of teachers. And then we just keep adding more to things like a coach where you can record your own lesson. And this is happening right now as well. You can record your own lesson and you can get feedback on the quality of your questioning, the quality of responses, how long you talk for.

And to practice that is powerful. I want to pick up on the point about the variability in teacher capabilities. Through no fault of teachers, like I was a secondary school maths teacher, you put me in grade one and I'm like, I have no idea what I'm supposed to be doing. And what do you mean children don't understand what five is?

We know a lot of our teachers in the global education space, perhaps haven't had the training, the right kind of training, the right kind of training. Background knowledge, so closing that gap is exciting. I worry about when we are struggling with teachers to even teach using pencil and paper traditional methods.

the demands of AI and teaching students to be more discerning, to understand and critique the information in front of them that's being produced, to have knowledge. We're still going to need knowledge because without that knowledge they can't be discerning and critiquing. The role of the teacher might drastically change.

And I'm not sure we are prepared yet to understand how the role of the teacher is going to change. And that when we have struggled to support teachers to teach literally the foundations of literacy and numeracy, how are we going to support them to have these higher order questioning skills. 

Pat: Yeah. So I see, I've been thinking about that as well, because we all know that one of the main purposes of these tools is to lighten the load and increase job satisfaction for teachers, let alone increase student learning.

So this actually speaks to the shared responsibility between developers and educators. So it's not only on educators to figure these things out. It's not only on education providers to train teachers, to use existing tools. It's on the tool developers themselves to create tools that are meant for education.

Right now, a lot of teachers are just using plain old vanilla chat GPT, which we know was not designed for educational use cases. And so there are AI and techs. Probably that Asyia that Gates is funding right now that are creating tools designed specifically for educators and so hopefully they're designing in a way that really reflects what teachers want.

Asyia: Developers, policy makers, funders. And Nissi, it would be great to hear from you about how are you integrating the technology with the pedagogy with the system? And what do we ask teachers to do less of if we're asking them to do something different? 

Nissi: Absolutely. Thanks, Asyia. And thanks for that, Pat, and just coming in there to say, It's part of why the work that we do, we advocate very strongly about the co creation and the co design aspect to building this.

Can we also bring this key stakeholders, the county, governments, you know, sub national governments and national governments, but also the teachers. And can we co create and implement pilots with this where we say, well, what is available on ground? You have two devices. How often, you know, how long are your classes?

What are the expectations that you have? What does the curriculum require you to deliver on? And so this ecosystem approach where we can bring all the stakeholders together and effectively provide that capacity and support to the edtech innovators. I think something else as well, and my final point on this is, Pilots and the roles that they play, because it's easy for us to sit together and, you know, co design and they implement this on the product.

But then if there are no pilots that give them the opportunity for testing and refining constantly until we're able to see that outcomes, then it still defeats the purpose essentially.

Ana: I think some very good points have been brought up by both Nisi and Pat, questions for both Nisi and Pat is. About half the world or a little bit over half still doesn't have access to Wi Fi access to the internet. They just don't have access to it. So we're over here iterating on rocket ships, but if we don't do the basics, I give them access.

Can we even talk about how revolutionary this technology is going to be? 

Nissi: I think that I can start to answer your question, Anna. I would say that's sort of what I wake up every morning thinking about, but I think that the biggest thing with this is. Intentionality in the design. We often see at CC Hub that we design with the realities in mind.

So we do not often just come with the utopia of, well, here's what's possible. Here's what we think can happen. And intentionality behind it and saying, Hey. We know that this particular number of people are possibly cut off. How do we design to bring them into the fold, even at the very, you know, basic level that sees opportunity for them to be able to derive value from these solutions?

It's what I said about the role of the pilots, because if you're going to implement a pilot as a test and refine cycle as an opportunity for builders, it brings everybody to the table and brings in that lens. In Nigeria, I was speaking to an internet service provider and they said to me, they said, Nissi, look, we have lots of schools, state governments who want to provide bandwidth to and want to provide it at an affordable.

However, if there are no solutions upon which they will be utilizing this bandwidth, then we don't see them renewing. We don't see them paying anymore because they just don't see the value. And they say, well, Nissi, we know CC Hub does a lot of this work in the ecosystem. Can you bring everybody together to figure out, okay, even if I have this amount of data that I can pay for per week or per month, how can we utilize that for them to still be able to get value?

So I think that to answer your question and not share this intentionality, but intentionality comes in two pillars. One is How we work with innovators and through the pilots and given a test and refined circle to cycle to build for the realities. And the second bit is with private sector and different stakeholders to then design models that see accessibility, even at the very basic levels, not necessarily at a utopia, but still see value in terms of access and use for the populations that we care about.

Pat: I actually wanted to ask Asyia, Asyia, you've been in this business of working with. Low and middle income countries for a while now, I'm wondering, building on what Nissi was talking about, what are the most innovative approaches that are really working, that are scaling out to large sections, because there are always pockets of innovation.

If you had one model to prop up and to scale, what would it be? 

Asyia: Asyia doesn't really have the answer, but I mean, one of the things that we know, and Justin Reich talks about this is even if you make it free. The poor don't necessarily benefit from it. So what Nissi said about intentionality is really, really important.

We have to hold ourselves accountable, and we have to prove that it's working for the most underserved. So one of the products we funded in India, they showed overall something like 0. 25 standard deviation impact, but for the furthest behind, it was 0. 44 standard deviation impact. And both those standard deviations are not sufficient because it doesn't get us, our children, to have meaningful literacy and numeracy.

10 children in sub Saharan Africa cannot read by the age of 10, and the mathematics themselves are similar, we need much more powerful solutions. And so, when we are funding, we have to ask ourselves, does the technology Do better than in the absence of technology and which kind of technology makes the biggest difference.

So, thinking about funding platforms like the SSA EdTech Fund, and this is referring to, is to bring together developers, government, teachers, and think about what's the most appropriate technology. And then have a measurement around it. But then another one is AI4Education. org by Fab Inc. And where we're thinking about what is the technical assistance global SAF developers need?

How do we support the southern developers? To have all of their talents, et cetera, be part of the solution by creating bespoke AI solutions. How do we make sure that the data sets are in local languages and the solutions have been produced for a wider range of people? And I think the other question is, is there an option to do nothing?

I don't think there is. The technology is developing so rapidly that if we do nothing, even in the presence of inequity in access, et cetera, the division is going to be larger. And one of the things I am quite proud of in the foundation is the AI task force, which is really thinking about ethics side, which is thinking about how do we understand the How developers in the Global South are thinking about applications.

So the task force ended up funding four Global South education innovations. And then we had such a rich array of applications that we then set up another platform to fund wider. But what we need to do is go back and layer the evidence and the impact information. Is it making a difference? Who is it making a difference for?

And it goes without saying, we're going to have an agenda decided for you.

Grace: Building on that, Asya, on your task force and sort of the intentionality of the global learning community that Nisi was talking about, something we've been thinking about at Atlassian Foundation is actually our work with our corporate partners. Atlassian. For those of you who don't know what we do at Atlassian, we're a tech company creating a suite of tools primarily known for project management and team collaboration.

But at our corporate and our foundation, we have a committee that we are calling the Responsible Tech Committee. And so two of the things that we've been thinking about is sort of the ethical framework and the oversight. And so what do you think is the Pivotal role of us in this room and other stakeholders to create those necessary guardrails to ensure ethical use and support of equitable access across regions and what's actually required to do this in that intentionality and collaboration and cooperation.

Pat: So, Grace, there are three primary areas of AI in education policy. One is governance, which I think is what you're talking about in terms of AI development. And then the other two are operations and pedagogical dimensions of AI and education policy. My work mostly focuses on the operational and pedagogical aspects.

Because the governance stuff is, frankly, in the end, it's going to be decided by governments. And obviously, we know that the EU AI Act It already got fully approved, and we know that there are some education provisions in that EU AI Act. There are two key high risk educational use cases that are regulated.

Any AI system that involves admissions to an education institution, and also any AI system involved in academic progression as well. And so those are two high risk use cases, and they go through a very strict regulatory process. In fact, teach AI just released policy resources recently, and one of the supplemental resources, if you actually go to teach AI dot org slash policy, for anyone listening to this, scroll all the way down to resources, you'll see a special resource called the EU AI Act memo, which is not only about the EU AI Act, it also talks about.

ramifications for regions outside of the EU and what we could learn from the EU AI Act. You know, the answer to your question, Grace, is I don't actually deal with a lot of the governance conversations. There are a lot of smart partners who are doing that type of work. Nissi, are you involved in any of the, the governance conversations?

Nissi: Again, as a work within the ecosystem to an extent, yes, and that's primarily providing insight just from one as technical advisors or support, but then just insight from teachers and also struggles that the innovators have. For instance, recently in Nigeria, there was the national AI, you know, strategy summit that brought together 120 AI experts, both of Nigerian descent within Nigeria and in the diaspora to just come together and decide and co create a strategy, Nigeria strategy.

On AI and ensuring that as a country, we're not left behind in this progress and development in AI and most times before it gets to the regulation bit is even just, can we have a set of standards or guidelines that innovators can build around or as various governments or even funders who are funding projects to look into?

But I think the biggest conversation within most of the countries is how do we ensure that The data that's being used to train it actually mirrors our unique context. And for cases where that doesn't happen, the data that exists, how can we ensure that there's that ethical access to it so that it can benefit the large populace as, as AI advances?

Asyia: Thanks, Nissi. I want to pick up that question about local context. I'm thinking seriously about when you have technologies and I know you alluded to this earlier about big actors working at a global scale can influence so much and then education being a deeply Local, national, subnational issue. What are the commonalities we're seeing?

What are the differences we're seeing? And is there a sort of a horizontal that applies across the board? For example, data protection, ethics, etc. And are there verticals that might apply in countries and regions? How should we think about this? 

Ana: It's a great question, Asia. At the local, national, regional level, we've actually made quite a few strides when it comes to policy regulation of these exponential technologies.

I'm not just talking about AI, although AI is obviously the big one right now. However, on the international level, transnational level, we're not there yet. In fact, there've been a few setbacks. So if we're looking at the world's You know, biggest international body like the United Nations for that kind of, so sweeping types of.

Regulation, not just regulation, but also what are the key things we all agree on? Like when it comes to the declaration of human rights, all nations got together and they said, we agree to these things. We all do. We're all going to try and be good at it. Those are the things we haven't come to an international agreement yet.

I know there's multiple efforts for them to be done, but we do not as nations of the world have come together yet to say we all agree the ethics should look this way. Accessibility should look this way. Regulation should look this way. It's still very fragmented process. And I think. It will come up to funders, it will come up to the stakeholders themselves to figure out a way how to engage with UN bodies and governments to think about this process a little bit, to come up with sort of a universal declaration of how we So Manage AI, the landscape is incredibly competitive and there's also huge opportunity here to make just a lot of money and so it's very hard and then there's also national security consideration so a lot of the nation see it as the next big frontier, the next big race and so that if they come to any sort of compromise or agreement agreement, they look at it as a possibility to miss out on developing, uh, or owning the space in a certain way.

And this is where the stakeholders can play a pretty big role in kind of directing. Unfortunately, from the U. S. perspective, you know, I'm based in Washington, D. C. Washington has invited all the big players into this. The stakeholders are still all the same. Guys that were, you know, part of the previous iteration of these technologies.

So the question then becomes, how do we bring the other stakeholders, the local regional ones? And, you know, how do we bring the teachers and the parents as part of these big conferences and discussions? In September, there will be the, the UN summit of the future. How does that become a good play, a ground, a place to kind of start these conversations?

Grace: Something that I have been curious about is if you can all weigh in on this. Will AI craft a new type of classroom with a new pedagogical approach, a new reality, and what do the young people and young learners need to be able to prepare for this type of classroom? 

Pat: Will AI craft a new classroom pedagogy?

No, AI won't do that. Teachers will do that, and AI will help a little bit. Changing teacher practice is very difficult, and you're combating multiple, multiple years of Teachers as students going through probably more than likely traditional ish, you know, models of education when they start teaching, then reality sets in terms of, like, the difficulty of employing more progressive advanced ideas and teaching methodology in, like, actual real classrooms.

And so theory hits practice and reality, and then, you know, theory goes out the window. So ultimately, you know, the one of the best things that might come out of AI. Two things. Number one, just the disruption to our just overall society, therefore causing education to rethink like what we're doing and why we're doing it.

And then second, maybe the tools themselves might actually help in small but significant ways. And then as for what skills will children need? What what's going to happen in the future is a de emphasis of some things and a re emphasis of other things. So for example, you know, problem solving skills has been a big push for decades.

But now that you have systems and tools that can do a lot of the problem solving for you, teachers and students are going to need to focus more on problem formulation. Like, why are we even solving this problem? Is the problem even worth solving? Which is going to be nice, because it's going to be more real world and more authentic, rather than just trying to like, get something done, asking ourselves, why are we doing this?

And what is exactly the thing that we're trying to get done? And another skill, skepticism and evaluation of anything. That's obviously something that we've been trying to build over decades as well, ever since the internet came out. And even more now when AI can create lots and lots of fake, untrue things, even more evaluation and being skeptical is useful.

Nissi: I think that to an extent I align with what Pat has said, right? I think one of the things that we're constantly thinking about is what's the role of teacher and how will the role of the teacher evolve when this tool is introduced and how do we support teachers to be able to do it more effectively because like Pat said, it's very important.

Difficult to change the habit or behavior that they've learned over a period of time. And there's just so much going on and that they need to do in the classroom and that they need to focus on. But we think that the teachers will innovate and all of this technology would support them to be able to think differently about their role.

And that delicate balance within, well, providing the knowledge, because like I said, we still need the knowledge anyways, but then also helping the children to build that higher Other soft skills, life skills, like you've described is thinking very critically about that digital literacy component and how are we upskilling them with that, but also ensuring that the tools are easy enough for them to be able to use and that learning curve, you know, is not as steep.

To see more effective adoption and value from, from those solutions. 

Asyia: This is such an interesting question. I feel like we should write our answers in an envelope and then pull it out in five years time, see how accurate we were. I completely agree when you see digital skills and what does it mean to have digital skills, especially with the advances in AI. But I think in terms of pedagogy, pedagogy is an interesting word and do we agree by the definition of it in itself? If you split it up into instructional practices versus why teachers do what they do, things will likely change. But I'm not sure we're going to see drastically different classrooms in five years time.

And teachers will take it up if it makes their lives easier, if it helps them to do what they're doing. It's defined as an arrival technology because people, teachers are finding it makes life easier when they, when they ask it questions and they can, you know, in response to them that way it helps them with their marking, etc.

But I think the use case for teachers is really, really interesting and I think also overcomes the access issue. You're more likely to have one device per teacher available than one device for students available, even if the student, one device was designed, I'm not sure it is. But I think the second use case that's really interesting is when we have.

A shortage of teacher skills and in our geographies when you're wanting to study computer science and you don't, you know, you're in a rural village or even in the urban area and you don't have teachers might not have the expertise, then solutions that can plug for that gap will be really interesting.

I think as an education sector, we need to be agile. I'm not sure we're a sector that is agile, but we need to be, and that's a muscle we're going to need. And thirdly, I agree with what Pat said about, it's going to be teachers who drive this, and teachers who learn from each other, and teachers who support each other.

They have a hugely difficult job to do, and I think the pandemic showed how important they were, not just on the academic side, but on the social side. And so an appreciation of the things that teachers do and that appreciation should be how do we support them to do things a little bit more differently and give them the time to and the resources to do that well will be key.

Pat: I had a question for you, Ana. You're dealing with diplomats all the time. I know AI regulation. Conversations probably come up, but AI in education. Is that a sector that diplomats are discussing? 

Ana: I would say not just diplomats, but policymakers. Yes, they discuss this all the time. And the viewpoint from the policy and diplomacy side of things is incredibly grim.

There's absolute misunderstanding about what the technology is doing or is promising. They see it from a national security point of view. They think algorithms and larger learning models are sinister in a way of saying. So when it comes to using them in the classroom, there's huge pushback on them. I would say some of it is grounded in good reasoning and some of it is grounded in sort of this dystopian, Outlook.

So it all comes down to huge misunderstanding. How big of a job your sector by your sector, I mean this, everyone represented in this virtual room today, you know, whether it's technology, education, philanthropy, your, you are part of this group of stakeholders that need to be able to better help with the understanding of what these technologies are doing to children, because when.

Regulation and policy does not understand what's going on. They're going to stifle innovation. And that's what we're looking at in terms of risks with this. 

Grace: I agree with all the things that, you know, Asya and Pat and Nisi have been speaking about this new classroom pedagogy. And I think the only thing that I'll just add to that is I really think that AI will actually enable more human centered and student centered pedagogies.

It'll really enable us to bring effective pedagogies in practice. I see the Socratic method really being powered by AI, where students are encouraged to think critically and adaptively through these AI driven questioning, tutoring, coaching, and support, which I think is that Continuous change and, and that move from sort of the rote memorization and, and lecturing and seeing students these like empty vessels that must be filled with the memorization of schooling.

And so I really think that AI can enable that.

I will actually jump to the question around sort of what can philanthropy be actually funding and something that I keep thinking about is what Nisi spoke to in the beginning, which is really about trying to do these iterative pilots and testing and refining. And so I think really where philanthropy has the best possible role in advancing technologies like AI is to essentially highlight that we could strategically be funding these high risk or Quote, quote, high risk funding opportunities that might be less attractive to government or private sector due to their extremely volatile nature.

As we were speaking, I was thinking of capacity building within stakeholders that are creating this technology, the infrastructure, like we were speaking about earlier, policy development and research, and then also building the knowledge ecosystem for AI. I think across. All of these places where philanthropy can really be strategic is collaboration.

So by working aside those in the classroom, talking to students in academia, in the industry, and even in government, I really think that can amplify that impact. I'm excited to kind of hear about your thoughts on your side. 

Asyia: So in the global education team, in the foundation, we have. A very specific focus, and it's how do we increase proportion of children to our foundational literacy and numeracy.

So the statistics that drives our work is if you're born in a high income country, 9 out of 10 children can read by the age of 10. If you're born in a low income country, 9 out of 10 children cannot. And similar kind of stats for literacy and numeracy. So what's the impact that we as philanthropists can have?

And I think there's three things. First one is quality. Making sure that when government is investing, it's It is on things that are most likely to have impact. And for that, we need standards, we need benchmarks, we need evaluation mechanisms that are comparable. The second was equity. Making sure that we are intentional about the underserved population.

And we're building tools that allows them to have access to high quality solutions that speak to their national curriculum are responsive to their experiences are in their languages. And allows them to accelerate and achieve what they want to achieve. And to do those, both those things, it's something you said, Grace, which is about coordination and collaboration.

And so that we have common tools for reporting on impact, we have common tools for really considering the cost. that there's no point funding solutions that are not going to be able to be affordable. Philanthropy can take the development and support the R& D to reduce costs, but that we are really thinking about how much would this cost a government to be able to fund and is it going to be affordable to them.

Those three things in tandem are really important and that we are in part of that coordination and collaboration. We're thinking about ecosystem side, we're thinking about the compute cost, etc. And my final point is that we shouldn't just think about access on its own. We have evidence that just hardware on its own or just having AI, it doesn't give impact on learning outcomes.

So that stack of high quality solutions. On a hardware that's affordable with offline, some kind of power and connectivity is going to be needed, but we're always thinking about what's the purpose of this and how does it improve outcomes for children, both on the cognitive side on literacy and numeracy, a wider understanding of how AI is going to be improved.

The future of the children who are in school right now, they're going to go into workplaces and they're going to have to be competing with people who will have expertise in this and how we make sure that we're intentional about supporting them to hit ground running. 

Grace: The curious question for you and Iosia, do funders and philanthropy understand tech enough? What should we be understanding in technology? 

Asyia: Oh, Grace, honestly, the best point I have right now in ChatGPT is like, explain it to me like I am five. The technology is so rapidly moving and it's humbling because it's made us all Learners, and it's made us all students. And then as soon as you kind of like, Oh, I know what random is now.

I know what fine tuning is. It's going to be something else the next day. Even the ones who are at the top end of understanding that technology, and I'm certainly nowhere near there, have to be rapidly moving with this. And that's why we need the experts to be able to speak to the lay person. And I think as funders, we need to think about what we don't know and how do we get to this.

So. Even this podcast series, hopefully, is helping to advance our knowledge a bit and understanding the application of those technologies in the teaching and learning process. Pat was talking about pedagogy and then understanding in the context where Nisi is working about how do we make it real and how do we make it impactful in a very diverse settings.

It's a lot to learn and we can only do that if we learn together. 

Grace: I think on our end at Atlassian Foundation, it would be quite foolish to think that we're experts in technology and all the other things that we're trying to really do and focus areas. I think I think that one of the things that Pat and Nissi and the Teach AI Initiative, this podcast, et cetera, can really help us just, and like you said, Asyia, inform us and help guide us.

At our foundation, particularly, we really strongly believe in partnering with the experts, and so therefore we collaborate really closely with experts like our grantee partners. Encode. org and to reiterate what Pat said earlier, really to listen and learn from them. 

Anjali: Thank you very much for listening in on the conversation.

Almost every week, there's a new event on AI in education and daily opinion pieces of webinars enthusiastically supporting its potential or urging caution. Funders like everyone else are striving to stay on top of these developments and asking important questions. Some of the key takeaways from this conversation has been that AI is revolutionizing education by supporting teachers across various contexts and offering benefits to students.

Despite associated risks, there are deliberate attempts to put in place ethical and governance systems globally and nationally. Funders should explore how AI can assist their beneficiaries. Preparing them for an increasingly AI driven future and as Asiya highlighted, while we can't predict all future needs of teachers, inaction is not an option.

This podcast was brought to you by the International Education Funders Group, curated and edited by Anjali Nambiar and post production by Sarah Miles. You can learn more about the IEFG at www. iefg. org and do subscribe to the podcast for more such thought provoking conversations.