00:00.69

Farheen Foad

Hello listeners, welcome back to Tech in EdTech. My name is Farheen Foad, and I'm going to be your host for today. I'm really excited for our conversation because not only will we be diving into all things tech and EdTech, but also getting a hint of non-for-profit sector within the education realm. I have a really special guest with me here today, Dan Sandhu, the CEO of Education Development Trust, here, based out in the UK. Hello, Dan.

00:39.35

 Dan

Hello, good morning. How are you?

00:42.36

Farheen Foad

I'm well. How are you?

00:44.21

 Dan

Good. Thank you very much for inviting me to this wonderful session. We're delighted to talk about education, the global development market, and generally the role of EdTech in what I've done in the past and what I'm doing at the moment.

01:01.29

Farheen Foad

Amazing. So I did, because most of our audience is based in the US, I did want you to do a proper introduction to kind of let people know what you're doing at Education Development Trust and what led you to come here so far.

01:17.21

Dan

Brilliant. No, I am the CEO of EDT. It's a global organization. We're committed to transforming lives through education and skills development. Our purpose is very simple is to increase the life chances of individuals by improving education and skills outcomes. I have been in education for the last nearly 15 years in one way or another, in everything from education technology, teacher development training, early learning, both as an investor, as an operator. My introduction to education came through as an angel investor in a number of education ventures, and I still remain an active angel investor in some education investors globally in the UK and in a number of international markets. During that period, I've worked in a number of organizations, been involved in listed organizations, as well as organizations which have grown significantly, both in curriculum-based education, which is facing off directly to students, but also in infrastructure education, which is supporting the environment, or in the assessment of technology space. I took over as CEO of EDT just over two years ago, and I was fascinated by the impact we could make at a global level with over 60 years of experience working with partners such as governments and foundations, et cetera. So I was very keen to take all my experience of working in private equity-backed organizations and work in a much more purpose-driven organization where we redeploy all our profits back into undertaking goods for the market.

03:10.62

Farheen Foad

That's amazing. And this is exactly the sphere of education that I myself am super interested in. So, it seems like ADT is a global organization, and as the leader of it, you guys are working from, my understanding, as far as rural Kenya to inner city England. So, as you think about AI's role in education, what's been on top of your mind this year?

03:39.73

Dan

AI is changing year on year. I initially got involved in AI over seven years ago in a machine learning-backed curriculum, facing a product in the UK. And I've seen AI evolve over that period. I sit on the UK's national think tank around AI and education on the advisory panel of that. And we're looking at how it benefits education from students, teachers, governors, parents, wider society. And what I've noticed, definitely in the last 12 months, is a rapid increase of adoption of AI in all forms, particularly generative AI. We have to be very careful because in international markets, the AI adoption is very different depending on the capabilities of what is possible within AI in some parts of the world, where technology is less easily available. But if just talk briefly about that one. We partner with ministries to support system capability building, policy development implementation, improving school quality, tackling teacher retention. And the key lessons I've got from the AI deployments from our work is about what you can do with using that technology to make a difference. Deploying any technology needs to be driven by defined and shared educational outcomes. So it's really important. That was one of our key drivers. There's got to be a shared outcome, or technology for technology's sake will fail, just as an AI is a subset of that. Any AI transformation projects that we've been involved in, we know need to be owned and driven by the Ministry of Education, by the stakeholders, by the school itself, rather than a third party, because that will allow it to have long-term sustainability. And it's also important right at the beginning of any discussions with the stakeholders, be they ministries or schools, or local governments, that the right policies are in place for all aspects of implementation, both ethical, safety, training, leadership, because the world is changing so quickly, the policies can't be static. And of course, the financial implications and implications on team members need to be considered.

06:14.60

Farheen Foad

Completely agree with a lot of the things that you were saying, kind of segmentizing and diving deeper into some of the things that you've said right now. AI deployment looks very different from context to context. So you might have a high connectivity school here in the UK, or you might be in a little resource kind of setting somewhere in Africa. So why is equity such a strategic focus when thinking about AI deployment?

06:42.39

Dan

I think the challenge has always been in education to get that balance right between equity and equality. We can talk about equity and equality in a number of different ways. The clear one, of course, is making sure we don't widen the gap between the haves and have-nots because AI requires particular technology. It requires particular bandwidth. It requires particular energy and resources in addition to the training and infrastructure. And we don't want to have a world where those with AI proceed at such a fast rate that those without AI get left behind. I think, however, this is an area where AI can have the greatest impact on driving educational change at the system level. Improving access to high-quality data about school quality and performance will reduce that gap.

And will support data-driven decision making. It also includes real-time monitoring and reporting at a national level and a sub-national level, which again then can form the targeted interventions. Even at a district level, AI can support improving things like school improvement and quality assurance because it will have access to school data, because it will provide that data to school readers and improvement coaches. So that will drive reduction in equity and inequality. But also, you know, it improves those individuals involved and supports coaches, giving them recommendations about interventions, about what's happening globally and locally. It can pull in a lot of data and support. So from a technical point of view, as long as we make sure there's access available, this will drive change. And the reality is the digital infrastructure and resourcing at individual school level will determine how successful this is. But this is a powerful tool if used in the right way.

08:51.49

Farheen Foad

You're sounding very cautious, optimistic, like most of the educators that I'm speaking to. And I think I couldn't agree more. It's just that I think we all have our own experiences that lead us up to these opinions that we have about AI and how it should be used. So my next question would be, was there like a personal or professional moment where you were convinced that AI could be a catalyst for equity, knowing that we set the bar right, that is?

09:21.23

Dan

I've been involved in education technology for a long time, for over 15 years, and I've run education technology in edtech organizations. So AI for me is a subset. It's just another technology. It's maybe all-encompassing, but you know much, much larger, but then so was the computers, mobile so was the whiteboard, so was the calculator. So we need to make sure we understand the significance of this technology in light of other technologies that have come into education. And the one thing that I was always very passionate about is to make sure that there is sufficient evidence of improvements, sufficient evidence of outcomes. Six or seven years ago, I set up something called the EdTech Evidence Group, which was exactly that, to support evidence and efficacy in the EdTech sector. And that was because I was frustrated at so many ed tech products which did not deliver outcomes. And you were actually, it was less about school wasting money. It was more about the edtech products wasting the time of the students and the children. And that really annoyed me. So I made a big fuss about efficacy. And that applies equally to AI. I know from a professional perspective that AI is a great catalyst to use technology to reduce that gap and make sure there's more equity. But I also, from a personal basis, I sponsor the adoption of AI internally within EDT. We're a multinational organization with operations in sub-Saharan Africa, in the GCC, and in Southeast Asia. So I'm very keen that we, given all our years of experience, use technology to the best of our capability internally. And I think that, and I obviously, like everyone else, use it personally every single day. And multiple times a day, we're using it as a tool to support, analyze, research, and I think it's exceptionally powerful. But, like every genie in the lamp, you've got to be careful what you ask. And you've got to be careful how you use it. So I'm cautious. And I think, you know, you need to be cautious. I may not be as doomsaying as some of the commentators, or maybe as critical as Geoffrey Hinton’s of the world. But I still think one has to be very cautious with all technologies, especially education technology, because it moulds the minds of children, and you've got to be very careful who you're delegating that responsibility to. So, AI with caution, as long as it doesn't undermine the teaching process or overwhelm the teachers and is there to support, I think it's a great catalyst for equity.

12:11.67

Farheen Foad

Right. I think it should be solving the existing problems and making teaching easier, learning faster in those kind of ways. But I'm curious because the context that EDT is diving in is a bit different from the conversations we're having here in the US. Some of the partners that you have are the ministries of education that are doing nationwide programs. So what have deployments of AI looked like in those kind of engagements? Has it been? Do you start from a small pilot program, or you go full scale with 50 schools? What is that engagement usually looking like?

12:51.99

Dan

I think we're in early stages, and I'm always very cautious when people use AI as a panacea for all solutions. It's not the case. I've been in large conferences, international conferences, where some ministers have funding challenges of a very basic nature, where AI is way down their list. So you've got to be very careful not having a product or a technology looking for a solution or a problem. The problems need to be solved rather than solutions need to find a problem. That's a really horrible way to run any technology business, let alone education. We work with ministries and partners, and schools around helping to develop education programs, system solutions, which include any support we can provide from an AI perspective. So my view would be this is, you know, and we use great partners who provide us AI support for data gathering in the background, the likes of Fab Inc in sub-Saharan Africa. So I think the aim is to make sure you find the right partners. But don't pretend to do stuff that you don't know because it will fail and make more damage than good. So when we're involved in AI solutions with our partners, with ministries, we make sure we have the right teams around the table, both internally and with the partners we choose. 

14:13.69

Farheen Foad

I think it also helps to look into the local ed tech scene that's happening within these countries. They're quite great at partnerships and working. They have the wits about them. I think it's just about scaling their solutions. And I think EDT and such organizations are great partners for them. I did want to ask, in your experience so far, given that maybe they don't have high bandwidth and one-on-one devices available, and there are bigger gaps within the education sectors, is there any specific AI feature that immediately helps low-connectivity and shared devices in schools that you might look towards?

14:56.74

Dan

I think AI works really well in a lot of environments. It's trying to figure out where the technology and the infrastructure is most suitable for it. If we are looking at areas where there's low bandwidth or less devices, it still provides a massive amount of data to drive quality in regions across nations or a multinational, particularly with international schools, for instance. But within the ministry context, it gives you a lot of data and information on which to act. So you don't necessarily need AI front and centre in the classroom if it's working in the background to provide you quality information about performance, about improvement, about school-level data which districts and school leaders could use to drive improvement.

15:49.42

Farheen Foad

So basically, using the data and the analytics more on an administrative level than like a student level, is it?

15:57.10

Dan

Yeah, I mean, exactly. I mean, and I know some schools where there's low bandwidth in sub-Saharan Africa have fewer devices in the classroom, but can be worked effectively. Remember, it's about the innovation of the teachers. A lot of this is about making sure your teachers are armed with what they can deliver and how they can best deliver it with limited technology. We're only a generation away from chalk and blackboard, so let's be very cautious about running before we can walk.

16:27.27

Farheen Foad

Right, right. Well, we've touched a little bit about equity and how AI solutions are deployed at scale globally, but I also wanted to balance the conversation a little bit by bringing forward some of the fears and concerns that education or educators generally have in regards to AI. So from my understanding, AI can surface powerful patterns and automate tasks, but at the same time, their automations can encode biases and even erode teacher judgment. So which teacher or student tasks look really tempting to automate, but you want to hold off on?

17:07.34

 Dan

So, as a general question, I don't specifically point this toward any organization, nor what we do at EDT. We're really careful when we talk about AI. We talk about AI, we talk about generative AI, you talk about large language models. Now let's really think about where they're being created. Which organizations globally are funding large language models, and where are they based? If you're in sub-Saharan Africa, how much of that large language model is based on African experience or sub-Saharan African experience? How much is it based on international experience, be it the US or China? You've got to be really careful what um what you drive out of AI in an education environment if the cultural biases built into AI are outside the nation. And that's up to the country making the call. I would be very disappointed if a ministry in sub-Saharan Africa chased the AI, went over the AI rainbow looking for the pot of gold, without really understanding where the source is from, what is a large language model, where it's created. So immediately worry about anything which is in front of the student, around not necessarily education pedagogy, but content, and context. So I would be very careful until those territories have their own large language models, until those territories are populating, as you know, making sure the learning aspect of AI is done locally. If the learning aspect of AI is done in Beijing and San Francisco, that's not really good for a kid growing up in the Congo or in Kenya, or in Rwanda. So I would balance that in those instances. Clearly, the best adoption of AI is the back office, where you're managing data and managing tasks rather than doing the teaching aspect. So we have to be very careful on what aspect of AI is used in which territory.

19:15.99

Farheen Foad

And from my understanding, EDT prides itself in the approach that it takes to education, A, that it's sustainable, B, it's also evidence-based. So I'm curious when it comes to some of the AI tools that teachers might be adopting, what is the process for trialing, planning the administration, and what kind of, maybe, numbers or metrics do you guys look at before scaling a solution essentially? Is there anything, any numbers or any tools that help you see evidence-backed solutions and see how it's impacting the learners?

19:55.53

Dan

So, from a technology deployment perspective, and I'll use AI as a subset, it's really good practice for any organization to have a good discipline and have evidence of outcomes. So the first challenge in any role, be it local, regional, or national, is understanding what good looks like. What is the outcome you're achieving, wanting to achieve, and have those metrics in place? Then have a period of time when you deploy that technology, and you can satisfy yourself before you use it in a sales pitch. You satisfy yourself, what you're doing in education makes a difference. Because if you can't do that, it's very difficult to persuade your clients and your partners that you are. Now, from an EDT perspective, we're very much involved in some regions in sub-Saharan Africa where we're doing teacher development training or supporting and employability. And in that respect, sometimes the active use of AI is much more in the back office rather than in the front office. But where we use AI, say in the UK, in some of the areas where we're doing career advisory, we clearly have personal intervention because in a lot of cases, when we use AI, we do have the personal intervention to make sure that the testing process is a real active process and the AI becomes an enabler rather than a standalone agent which does A to Z. We've got A to Z, we've got to get the balance right. I think we're very good to make sure we take a very teacher-centric or advisor-centric approach.

21:32.55

Farheen Foad

We've seen so many international edtech partnerships follow a pattern where they do a pilot, they're doing external funding, showing promise of results. But as soon as funding ends, everything stops. And I think one thing I highlighted is that EDT has a very sustainable approach to the solutions that they put, whether that's in the UK or globally. My question around that is how should ministries and other schools, vendors, or even NGOs like yourself create mentoring networks, ownership models, so that collaborations are essentially outliving the pilots and truly becoming sustainable?

22:26.32

 Dan

Absolutely true. So just to be clear, we're not an NGO, we're a for-purpose organization. So I know there's a lot of large NGOs like the UNICEFs of the world, which we work really closely with, we're much more in the for-purpose, not-for-profit sector. I think for us, it's about working with ministries to understand what they need and what they require. There's a lot of organizations who rightly say put AI on their presentation deck so they can raise money. And when the money runs out, they have no solution. Really poor business model landing, and it never generally works. The aim always has to be that you're delivering something which makes a tangible outcome, and you are using much more interest in the metrics of educational outcomes rather than metrics of education, but technology outcomes. We're very keen; I've always personally been very keen to focus on the educational outcomes which make a difference, and therefore that drives efficacy and efficiency in the technology you use, and we try to get that message over where we use AI internally and where we support any products externally that we focus on the outcomes and they're very clear to us.

23:41.05

Farheen Foad

And is there any example of maybe like a gamified STEM challenge or AI-enabled mentoring program, as you mentioned, that survived major context shifts?

23:52.74

 Dan

AI is too young to have any valid case studies. I remember in previous organizations, we ran multi-year, multi-tens of thousands of students using technology products before we were satisfied they gave relevant results. The AI products available last year have almost become redundant now. So I think that's going to be very difficult to get any solid case. People, students, teachers, schools, districts, regions, countries are going to have to look at what they view as the step change has been. And I think that's not easy, and it's not about just saving money, but making sure you're really clear about the impact. I don't think there's been valid user testing, or at least sustainable long-term testing of products with an AI bent, which would have made a difference yet. I think it'll take a while to get long-term, sustainable testing.

24:58.68

Farheen Foad

All right. We're nearing the end of the podcast and, not to put you on the spot, but I've prepared a quick set of lightning round questions.

25:09.34

 Dan

Oh, God, okay.

25:09.34

Farheen Foad

Are you up for it? It's a one-word answer, maybe a longer answer. It's completely up to you. But if there is any one AI application for equity, you'd prioritize funding this year, which one would it be?

25:27.55

 Dan

From a funding perspective, I'd be very interested in anything which drives more information available to the ministry at a national level so they can make better decisions.

25:40.40

Farheen Foad

Amazing. One automation and education you'd never green-light, no matter how much time it promises to save.

25:47.86

 Dan

A robot stood in front of a classroom.

25:52.33

Farheen Foad

And if you could only track one metric to know whether an AI tool is genuinely supporting students or teachers, not just adding to their work, what would it be?

26:06.09

 Dan

That's a really good question. Depending on where I use the product, front office or back office, if it was front office, I'd be looking at a metric which demonstrates long-term retention of learning rather than short-term retention learning. If it was back office, I'd be looking at the efficiency of tasks undertaken by a teacher, particularly, which takes them more into the classroom and out of admin.

26:30.81

Farheen Foad

All right. And last question. If you could leave education leaders with just one principle to guide their AI decisions, something that would help them deploy tools that truly narrow the gaps, what would it be?

26:49.31

 Dan

The one advice I would give teachers and school leaders would be, you're the teacher, you're the school leader. This is just a technology product. Don't get bamboozled by the technology. Don't get bamboozled by the terminology. Focus on the pedagogical delivery of what you are good at. And therefore, challenge anyone telling you there's this solution which is going to save you time or teach your children better, or give you better outcomes. Challenge, challenge, challenge until you're satisfied. Not necessarily value for money, but value for time that you're investing in this to deploy and see the outcome. Do not trust anyone selling you anything until you see the evidence. And that is going to be a tough message for AI providers and technology providers. But if they do not have the evidence, sustainable long-term evidence of success, they should not be in this business.

27:45.22

Farheen Foad

I couldn't agree more. I think educators and teachers are really the heart of our whole school structures, and without them, there is no moving forward. Even with new technology and developments, there will always be a need for a good teacher, a good somebody who can come across to the students. The human in the loop will always be a big factor.

28:08.44

 Dan

Absolutely. We all remember our best teachers. However many years ago, very few people remember other people in their lives. But you can always look back 5, 10, 20, 30 years and remember your great teachers. There is a reason for that. You're not going to remember the name of the AI app or the technology you used, but you will remember the name of your teacher. And there's something to be said about that.

28:33.03

Farheen Foad

I agree. Thank you so much, Dan. I really enjoyed our chat, and we may not be able to close all the gaps right away, but talking about them and really recognizing the inequalities is where the change begins. So thank you so much for joining us today.

28:45.79

Dan

It’s really my pleasure. Thank you.