Cameron Mirza and the Learning Scientists

Some university leaders believe AI is free.

Not cheap. Free.

That's one of the findings from research by our special guest this week — Cameron Mirza, who leads major education projects in Jordan for IREX, the NGO working in over 100 countries, and who has just surveyed university AI readiness across 25 countries. Three-quarters of the respondents came from the Middle East and Africa, a perspective you rarely hear in the UK AI conversation.

The headline numbers deserve a moment:

Only 34% of institutions have a clear AI strategy. Only 39% have approved AI policies. Universities have grand AI ambitions — Cameron hears them in every president's office, complete with impressive PowerPoints — but ask a simple question like "can you give me an example of how you've used it?" and the fundamentals fall away. No strategy. No budget. No staff literacy. And in some cases, a genuine belief that the whole thing costs nothing.

His diagnosis will sound familiar to anyone who's listened to this podcast: ambition driven by fear of missing out, budgets that don't match the rhetoric, and a rush to buy tools before anyone has worked out what problem they're solving.

But the conversation goes much further than the survey. We get into why the Fable ban has made AI sovereignty the topic in every university leadership meeting from Nairobi to Abu Dhabi. Why offline LLMs and low-bandwidth design might matter more for global education equity than any frontier model. The Stanford research showing students achieve far more with a teacher plus AI than with AI alone. And Cameron's framing of the future of the profession, which Neil announced on air he was stealing: teachers as learning scientists — part educator, part data analyst, part behavioural scientist.

Oh — and the story of how Cameron once collected four months of unpaid salary in cash, in carrier bags he had to fetch from the car, then hid it in suit pockets and under his mattress because the bank wouldn't take it all at once. You'll have to listen for the full glory of that one.

We've known Cam for over 25 years, since the three of us fought the millennium bug together in Ford and then went on to the Department for Education. This one was a genuine pleasure.

Two mates and an old friend. A bar. And all they want to talk about is AI.

Pull up a stool — we'll get the beers in. 🍺

#AI #HigherEducation #AIinEducation #AISovereignty #EdTech #InternationalDevelopment #Podcast #AIStrategy #FutureOfWork

Transcript

SPEAKER_03 0:00

Are we good to go, gentlemen? Uh should we get this pantomime horse of a podcast under this six-legged uh pantomime horse of a podcast underwear? Because this week we are very blessed to have with us an old friend. And I say old, what I mean is he's not very old, but he's been a friend for a long time. And he just reminded me it's over 25 years, which is absolutely amazing. So silver jubilee, that means. So welcome to Cameron Mercer. Uh, for those of you don't know, Cameron is a really interesting chap. Uh, we met, as he said, over 25 years ago. This time, 20 years ago, we were working in the Department for Education, all three of us, doing different programs and different things. And um, Cameron uh now works and lives in the Middle East, and he's doing some really interesting things with governments, with universities, and most importantly, from our perspective, with AI. So, so Cameron, very welcome to you. And would you, for our audience, uh, if he's still awake at this point in time, would you just give give him a little bit of background uh on you and what you're doing and and where you find yourself in the world right now?

SPEAKER_00 1:22

Well, first of all, Neil and Kieran, it's great to see you guys. As as you said, Neil, we have known each other uh a long, long time, and it was our 25-year anniversary. I think it was last year. I think we all first met in the year 2000. Yeah, I think I think that's where actually when we first met. Um when we were fighting the millennium bug. That's right, that's right. It was, yeah. Y2K successfully we we thought we conquered it. Yeah, we did. We conquered it. I remember the Y2K project, yeah, year 2000, and uh and uh they were they were definitely great days, and um, and then obviously uh we all then uh worked at the Department for Education, um, which was you know a really interesting uh period of time. It was it was the early 2000s. I think there was a general positivity about the country. Um, I think Blair was prime minister at the time, and and and that things were really moving, I felt. I felt I felt that education was a priority, it was really moving, there was significant levels of investment, and I think we were privileged, I think, looking back to work on some really cutting-edge uh pieces of work uh at that time. And um and I think we did we did uh make make a difference. I I think ultimately I I really feel looking back, actually, perhaps um we should be we should feel really proud of some of the work and some of the projects we worked in at the time. I remember London Challenge was a was a great was a great initiative. Um I remember some of the stuff we did uh for uh youth um um uh youth poverty was was really interesting work. I think some of the work we did in terms of digitalizing the uh the department's um data, you know, with our colleague and friend Kevin at the time who did incredible work with with uh local areas and drilling down into data. And I think that was probably we were ahead of our time really in terms of data analytics, um, and and some of the fantastic work. And uh, you know, I obviously had the great uh pleasure uh working with um uh Baroness Louise Casey on a couple of occasions, which was a very which was a very interesting experience. I can tell, I can tell you, and probably our friend Ashley probably still hasn't recovered from that experience.

SPEAKER_01 3:52

So um no formidable force, she was, wasn't it?

SPEAKER_00 3:57

Yeah, she is, she still is, she still is a formidable force. Um so yeah, I mean it was amazing, it was an amazing period. And then I think um I I do think I do think I think it was down to another old colleague of ours, I think at the time it was Janice Shiner. And um she Janice, if you're listening. I don't think she'll be listening, Kieran, somehow. I don't think so. No, I don't think so. But better things to do. Um, but I do remember she was uh I think she was a DG at the time, and I think I think she played an important role in the next phase of my career, really, because I think they were she was doing some work in Bahrain at the time as a as uh an advisor to the government, and I and I do believe that she put me forward for uh for a job there. So um yeah, she did, yeah. And so so um I remember it was uh I think it was about 2011, and I remember my phone ringing, and uh it was a number, it was in this national number, I had no clue where it was from, but but I answered it, and it was the chap from Bahrain saying, Listen, um you've been recommended, and um, would you be interested to come and work for us in Bahrain? Uh we're doing we're kind of re-vamping the higher education sector and um looking for somebody to come and help us. And um uh I looked out the window and it was uh pouring down with rain, and I looked at the weather, I looked at the weather in Bahrain and I thought, yeah. And um when I spoke to I spoke to my good friend uh who you've met many times, Pat, who's a tax expert, and he said, Yeah, Bahrain's a tax-free country. So my mind was made up, and so there I went. As simple as that.

SPEAKER_01 5:44

Wasn't there when you first got there, there was a problem with you getting paid promptly enough, wasn't there? Do you want to feel you can share that with our audience?

SPEAKER_00 5:52

I think I can long enough a girl. Yeah, I think I can share it. So I mean, I was a bit uh maybe naivety at the time, or or you know, just simply trust, really, because you're dealing with the government. So I went I went to Bahrain with one suitcase, and um I had no clue what I was walking into, I'd never been there before. And so um I I kept saying on the Skype at the time, I was saying, well, you know, what about the contract? What about did you you know, I must sign a contract? What I'm I didn't even know what I was getting paid, to be honest. Um, and so I I went out there with a suitcase, and um, to be fair, the first day I was there, he put the contract in front of me and he said, There you go, and got it all signed. Uh, but the bottom line is I didn't get paid for four months because of their systems, and um, you know, uh moving moving a contract about requires several stamps on it in those days. In fact, in fact, still now the Middle East and many parts of the world still operates on this paper-based system and stamping things. And so my my contract got shuffled about from department to department, and um, and by the end of you know, the third month, I I had no money left. I was running on fumes, and uh, you know, I'd taken an apartment, I'd taken a car, and so I just said, Listen, I'm being paid, uh, I've got no money left, and I've I've really got to go back unless you pay me. And my boss, um, who's a very dear friend of mine, um, still, um, said, Look, I'll sort this out today. And I remember being driven to um a facility somewhere, and uh I got there and the guy, you know, said, Have you got your passport and so on, handed it all over, and he said, Okay, have you got a bag? I said, No, I don't have a bag. So I went I went back to the car and there were there were there were some um you know carrier bags, so I took them in and he and he basically counted out my my almost four months' salary in cash. He put it all on the on the on the table. So I love it. So so I said to him, Well, what am I supposed to do? Well, I said, What am I supposed to do with this? He said, Well, take it to the bank. And I took it to the bank, and the bank couldn't accept it because they said, you know, there's a there's a threshold for accepting cash payments in a day per person. Well, that's it. I had to go each day and I had to go back to the apartment and hide the cash inside suit pocket and under the mattress because you know, the where I was living at the time, they they'd have a service where a cleaner would come in every day, and I I thought, well, I can't leave his cash lying about, obviously. So I literally had to hide it until it pegged in.

SPEAKER_01 8:44

I like it. I love the direct direct solution. That's brilliant, isn't it? This thing's been on so long, we're gonna fix it right now. What a great uh great attitude that we have. But bring your old carrier bag.

SPEAKER_03 8:56

Bring your own cash carrier bag. I love it, it's a very cool and tell us, Cameron, what are you doing now? Move on to move on to modern times. What are you doing now?

SPEAKER_00 9:05

Yeah, modern times, modern times, in in 2020, it was during COVID actually, um Irex um uh uh an NGO um based out of Washington, but you know, working working um in over 100 countries, and a lot of a lot of the work is really based around lower lower middle no low middle income countries. And I think as you said earlier, you know, in one of your podcasts, I think uh recently about you know serving the underserved, and a lot of our work is serving the underserved. And um in 2020 they um they had a project which had just started in Jordan, which was a pro a project around uh uh pre-service teacher education, training up the next generation of teachers, and um I think because of COVID um they couldn't find anyone to to to sort of run the project, and um and so to be honest, they took a bit a bit of a huge punt on me because although I had the kind of educational side of it, I hadn't worked in development, I'd never worked for an NGO. I I was totally new to the sector. And I think I think out of pure desperation because they didn't have a plan B, they gave me the they gave me the opportunity. Uh and actually that's that's actually true. They didn't have a plan B, and I was the only option available on the table. So they said, well, we'll give it we'll have to have no option. Reluctantly, Mr. Murder.

unknown 10:34

Yeah, that's right.

SPEAKER_00 10:38

So um, so it was again it was a case of I'd um I literally came with a suitcase because the you know, the time the time of COVID, as you remember, borders were shutting, it was it was almost impossible to get in out of countries. So literally it was a last-stitched effort to get into the country, and um, and because of COVID, I couldn't leave the country for for nine months. So it was uh so it was a challenging situation um at that time, but um uh but you know, six over six years later, I've just I've just come up to my six-year um mark with um with IREX. And you know, we did a we did a a pretty good job with that project, and um, and then uh you know in 2023 I was uh privileged enough to be asked to run a new a new project that they'd won, which um was was really focused on early grades education, and um, you know, I I was uh I really was delighted to to accept um that you know taking on and leading that project. So yeah, so I've run I've run two major projects um in Jordan uh for IRX, and as well as that I've been kind of working with colleagues uh on leading on AI, which we've really focused on in particular over the last 12 months. That's become a real area of focus for us. But um, but in terms of the work here, we've actually been working on AI for a couple of years now because we've seen the potential and the opportunities, and so and so you know, I feel I feel like we're we've been ahead of the game in some in some aspects in terms of uh AI.

SPEAKER_01 12:11

Yeah.

SPEAKER_00 12:12

So that that that brings us up to today.

SPEAKER_01 12:14

Yeah, and you've been doing research in this area. I know IREX have put out some research, haven't they, across me now, or probably wider than now, I imagine, universities particularly. So obviously that's an area we're very interested in, AI and education is uh big focus for us. We and we work with universities. And I I read you had an article in get this University World News, the only publication with uh maybe even less less readers than we have listeners. It's probably massively popular. It's just my ignorance that I hadn't found it before uh before looking on this.

SPEAKER_00 12:53

Yeah, it's uh it's uh you know, it's for those in the know, Kira, is how I would describe it.

SPEAKER_01 13:02

Perfect. Well, you said so some stats from there that I'd love to just explore with you is um only 34% of the institutions surveyed have a clear AI strategy, which is really interesting. Given it's an AI survey, presumably, although that you can tell us a bit more about it, then to answer and still not have one is interesting. And that only only 39% had uh approved AI related policies. So we're running here according to the National Education Union's 2026 survey, we've got 40, was it 49%, 48% of schools have an AI policy? So still not enough, but getting there. Um that's this is a lower than that. And I guess in universities as well, given the kind of the cheating AI for cheating your way through a degree, I guess having a policy is arguably much more important.

SPEAKER_00 13:53

Absolutely. I think I think what brought us to this work, um Kieran, was uh we we we work with universities here in Jordan, but you know, we also work with universities around the world. I RX has got affiliations with hundreds of universities, um, and and do a lot of work, particularly on the kind of um exchanges, and particularly with US universities, where where obviously we've got a strong presence. And um I'll often speak to presidents at universities and they'd often tell me about their AI ambitions, and they wheel out these really you know uh grand PowerPoint presentations about how they're planning to use AI and how they're using it, and then I would ask some simple questions like, Well, can you give me a an example of how how you've leveraged AI? Um what you know, do you have do you have you know fundamentals really? And we describe them as AI fundamentals in terms of the policies, in terms of the strategy, in terms of the capacity building, in terms of the financing, in terms of the the digital infrastructure, really just the fundamentals. And me and my colleague uh Tom Oral, um we spent a lot of time thinking about this, and and you know, we came to the conclusion about a year ago that universities have got great ambitions, and I think some of that is driven by noise, some of that is driven by fear of missing out syndrome, um, and and they feel that they need to be pushing on AI. And I think they're probably right, they need to be, um, but but really lack the fundamentals of how how you go about it. So um so we really became curious, and we we we designed this kind of survey which reached 25 countries in the end um to get feedback on on how they're using AI and and some of the kind of key questions that um will help us to understand what the current status is of AI. And 75% of our respondents are actually from Middle East and Africa. And and so you know, the results that came back are, you know, broadly speaking, yeah, uh universities have got great ambitions to for AI, uh, but but lack AI leadership, um, lack lack the strategy, lack the policies, lack the um, in many cases lack the digital infrastructure, and for us importantly, lack the um the the um AI literacy amongst their amongst their uh faculty and staff to leverage AI. So there are huge gaps in all of this. Financing is a massive issue because um you know when we go and talk to universities, I mean we've just come back from recently Nairobi, where we met with 25 universities in Kenya, and obviously we we've had a lot of dialogue with universities in the Middle East region. There are there are uh universities and and vice presidents who believe that AI is free. They do not they do not believe they do not believe that they're paying for it. So when we have these conversations that you know it's not free, they're a bit taken aback, and they you know, there isn't that the financial planning to understand, you know, um the price of tokens. Well, I mean, token prices are gonna you'll know better than me, but they're gonna go up for sure in the next few years. And has that been factored in? Vendor lock-in as well is another issue that comes up quite often. Um, cybersecurity is another area. So there is this, there is this kind of area of holistic assessment of of institutions to understand where are they in terms of their AR readiness and and putting in uh practical steps to help them get to where they want to go, really. So I think that's really where we're at at the moment in terms of that, in terms of that work and and and the data we collected. Nice.

SPEAKER_01 17:37

It's really good that you're kind of driving that conversation. I'm really pleased to hear that at least they have grand ambitions, even if perhaps sometimes the fundamentals aren't yet there, but at least starting with a vision and then gives you something, doesn't it? I saw in that report you mentioned finances, but um um uh 27% of institutions have dedicated budgets for AI. This is one of the things that really I when I'm on stage talking about AI and education or indeed other public services, but I get kind of frustrated about because I think I think now most people, arguably everybody, get that AI is going to be pretty transformational and as I always say, it could be made for education, the ability to personalize and you know meet different learning requirements and neurodiverse needs and help are targeted for for those more um perhaps underprivileged learners that might need it in a different way. I think everybody's I think gets that, and yet when we'll talk to an institution over here, they might they don't have a budget for AI and they'll be scrambling down the back of the settee for you know £11.32. And like you said, they think it's kind of free. And um, I I find it so frustrating, but it's I mean it's kind of a good that third, I mean nearly a third have a budget, everybody needs one. Do you see that? I mean, I because it is changing by definition, but do you you think if you ask this next year we're gonna see 100%?

SPEAKER_00 19:04

I don't I think we'll see an improvement, Kieran. I think I think there has to be, I think, I think um if to meet the ambition, there has to be um budget. It might be it might be disguised under or under the the term digital transformation, for example, but but essentially I think I think budgets have got to have got to align with ambition if they're if they're gonna um create sustainable impact from AI. Um but you know here's the dilemma. Many institutions, uh particularly parts of the world that we work in, have got very limited budgets. And and that's really that is that is really the challenge here. And I think um there are many I mean I think the universities um there is this kind of I feel a bit of a a crisis heading their way. We've seen it in the UK with universities looking at merging. We know some universities have been bailed out by government uh by government uh over the last few years. Yeah, and I and I think looking at many parts of the world, whilst the demographics, particularly in places like Africa and the Middle East, show more and more young people entering higher education, I think there are different choices available now for young people. And that the traditional you route of of universities, for example, might not be as appealing as it was a few years ago, um, and there's a lot of competition. Um so I think I think I think universities face a couple of crisis points. One is how to use AI, and yeah, and I think the other one is is is actually how to survive. Because I think it's becoming increasingly competitive out there. Students have got a different way of the world, you know, uh a different outlook. Um, depending on where you are. I mean, you know, I read something recently about the the the debt in US for the student debt for for for colleges, uh something like $1.7 trillion the debt for students. And at the end of the at the end, at the end of that, you know, you really want a job, don't you? You want you want to you you really want to find a job that pays you well and and you know allows you to get on with get on in life. And I think um so I think priorities are shifting. Um I think definitely universities have got to look at AI not just in terms of teaching and learning, but for a lot of the administrative and back office aspects. Um, how do they recruit students? How do they use AI for really, really personalized recruitment?

SPEAKER_02 21:33

Yeah.

SPEAKER_00 21:33

Um and you know, as you guys know better than me, how do you automate a lot of the workflows? How do you become more efficient? And I think those conversations um are going to start to happen um sooner rather than later. Yeah.

SPEAKER_03 21:47

I think student debt in the UK is is a massive problem too. And it can't just be the UK and and the US, but yeah, certainly here, lots of people are questioning the kind of should we actually go to university because I'm gonna end up with 70 or 80,000 pounds worth of debt and I'm never going to be able to pay it back. So yeah, I think it is a I think that how universities are funded and how they survive is going to be a real challenge in the UK. But I'm assuming that's the that's the case for for other other organisations well as well. So one of the questions I had for you though, Cam is um because you're because you're uh Mr. World Traveller these days and you're off around talking to all these universities, is there a particular institution or have you seen uh institutions where they're using AI well and you would think would be a good exemplar for others?

SPEAKER_00 22:36

There are a couple actually it's a good question. I mean I think we partner with Florida State University and they are they are really embracing AI you know from end to end you know from from student recruitment uh all the way in through the teaching and learning to you know to to to a great example they gave they gave us recently was how they're using AI for allowing students to do mock interviews for for jobs. So they get you know which is great isn't it I think you know a student would would would do a mock interview and they get instant feedback on how they did and I think I think that's really that's really clever how they're using that.

SPEAKER_01 23:19

Arizona State are using a lot of AI and they've got a lot of um courses on AI um but you know they're they're they're really uh they're really kind of proactive in terms of how they use ai I think also I think um in the Middle East there's uh in UAE Mohammed bin Zayed uh university for AI and so there's a specialized university just for AI based in UAE and they're doing they're doing a lot of um cutting edge work um and they're developing their own um llms in arabic so I think I think what what you're seeing in in in certain parts of the world is is institutions at the forefront of of AI sovereignty and I think I think certain countries are starting to really move in that direction and sovereignty is becoming a an important part of the conversation now um particularly particularly when I'm when I'm talking to uh university presidents so there's definitely a shift yeah yeah and with kind of the restrictions on uh the anthropic's fable model that they was pulled under White House requirements and 5.6 chat GPT 5.6 was uh went through a bunch of checks before they let it out I think that that bit where they pulled fable the anthropic model uh from everybody is I think was for me a very big moment in understanding quite how dependent we are on the US firms for this stuff and what that could mean should we you know I mean who knows what what can happen in the world of politics it seems to change quite frequently doesn't it but I think sovereign large language models particularly I think is going to be really required. It's hard to imagine isn't it that you could have someone just switch off all of your AI capacity particularly you know we think internet is now kind of absolutely ubiquitous to running everything from traffic lights to water to hospitals um you know AI is going to get weaved into all of that pretty rapidly and the idea that someone could just pull the actual model from underneath you and your services is very concerning. So and I I think the Middle East has always been quite big on sovereignty I've done a bunch of work with banking through the years and I remember the Middle East is one of the most restricted for saying you cannot let the data out of the Middle East on anything and the regulations here and no but you can have you can have three carrier bags worth of cash though.

SPEAKER_00 25:49

It's so safe well I mean a great example of that Kieran when I was at the University of Bahrain we actually did a deal with um AWS and they built a data center on the campus. Wow yeah and it I mean I think I think there was a win-win there because yeah they built a room yeah yeah that's right and um and our our our students would get jobs with AWS which was a which was a great which was a great win win but you know but definitely if you look at Saudi you know the investments they're making in data centres is is unbelievable in you know the level of investments but you know the other side of it which we don't we perhaps we don't talk about enough is the impact on the environment the amount of water needed to cool these data centers is is just unbelievable.

SPEAKER_03 26:34

So there are you know there is a there is a trade off to all of this but yeah you're right you're right near uh definitely the Middle East is really um uh increasing so uh and also because of the geo geopolitics particularly over the last uh period is it's pushing more and more focus on on on sovereignty yeah the geo the geopolitics thing is interesting to me because um you've talked kind of at the high end of education about universities have you seen any examples on your travels where there are the kind of at the lower end so you know in schools or where you know uh education is is a real challenge because of uh schools electricity yeah yeah big classroom sizes small teacher numbers you know people where countries where or or regions where they're struggling to deliver basic education have you seen ai filter down to any of that stuff yet or is it too sync?

SPEAKER_00 27:29

Well I've seen I've seen some I've seen some um interesting case studies uh Neil I think that the opportunity is for low bandwidth solutions and I think you know smartphones almost even wherever you are you know smartphones is is the obvious uh option there and I think the other the other area which will emerge is offline LLMs I think will be a will be an opportunity as well so I do think whilst um you know the I think fragmentation is is a is a challenge based on where you are fragmentation of AI usage because of the infrastructure I think that more and more um philanthropic organisations um donors um are looking at these kind of solutions particularly in Africa where I think there's a lot of investment going into Africa particularly I think the Gates Foundation for example is a is a great example but but definitely there'll be a big push on these kind of low bandwidth and off offline LLM solutions over the next uh period and I think that's an exciting opportunity you know one of the things that has really sort of you know bound us over the over 25 years Neil has been our our passion for serving the underserved and I think that's really you know if we can do that through our work particularly um the work I'm leading on here in Jordan I think that'd be you know that's that's legacy work for me. You know and I think I think we've got to and and and and it's also helped me focus our attention on on on on shifting towards um more integration of AI in our work. Because I think your point uh Kieran was absolutely right you know the cat's out of the bag right it's flying and we've we've just got to help countries and help school systems um catch up quickly because the worry I have is the the equity gap gets bigger and bigger between those who have and those who who do not have so you know that that's really you know uh driving our work on a day-to-day basis.

SPEAKER_01 29:35

Yeah and and as we've talked about before on here is the free tools available to people arguably who are less well off are the sort of equivalent of having your sort of £2030 accountant versus your really expensive some of the really expensive pay poor tools equivalent having a £100 120,000 pound a year accountant and your you know the difference in what you'll get is massive and I worry on a more colloquial level really AI detection I think if you're clever and you have decent models you can easily evade AI detectors. If you're not so you're gonna get caught up in those so I think there's a whole bunch of sort of things that that catch you out. We build knowledge flow interestingly there's a t a setting on how many times it will kind of wait and recall on the network and that's all about your internet speed and reliability and we took we push those quite high on knowledge flow which does occasionally mean you're delayed longer than you need to be but it means that if you are living in an area with rural Wi-Fi or you know you're dialing up on your phone or whatever, then um then you've got um you can still use the tools and they'll they'll still work for you. But yeah who knew things like that actually matter um and tiny decisions you make can make a very big difference and part of the reason I like the idea of sort of sovereign models universities and institutions creating their own thing is you can make it equitable for all of your students and staff um which doesn't stop people from still obviously dabbling in the more expensive stuff but at least you're bringing the bottom up and raising raising the floor which is you know at least doing something so it's uh I think that's a really good point here just on that I mean one of the one of the bits of work we're doing right now is a is an AI sandbox with with um a university here and and and and the reason behind that is uh we we are working with them on using open source open source tools to to curate them and to test them in a in a safe environment both for faculty and students are are developing and testing these tools um um because I think I think uh what we don't want them to do is as we've seen often uh spend a huge amount of money on on on buying tools um scaling them up and then getting no traction so I think we are we are taking this kind of more pragmatic approach yeah good uh through through this change everything I mean it's just as we continue to say as everyone says but as we continue to say uh AI implementation is not a technology problem it is a culture change management all of our world that we've lived in for ages is about getting people on the journey with you and I interestingly my perception from my uh knockings around with our customers and clients is university staff are some of the most resistant to you know they they come with a lot of uh concerns about using AI and I can kind of get it academic discipline and you know the stuff that they live and breathe is you know easy to shortcut let's put it that way uh and not not rigorous or not as rigorous as as they would like but it is a real challenge is to sort of try and win the argument a bit with well you know you don't need to give it everything we're not asking this to be the professor or to do the research but look what you can enable to take away some of that you know the heavy lifting part that frankly right you're right you know it isn't really a job for a a deep professor researcher you can you can give that to some AI and then spend your time on the stuff you really want to do.

SPEAKER_00 33:11

You've got to win an adoption I think I think on the adoption what we've done with the university you're absolutely right the the resistance to change is is is you know is is incredibly high uh the way we've approached it is by applied training so they're actually not it's not theoretical we're getting them to use use tools hands on to see the impact very quickly but also identifying user cases very quickly as well and and and and amplifying those user cases they guess what you can achieve but by by doing this and I think that that that has helped us break down some of this resistance.

SPEAKER_03 33:46

Yeah good I think if you look at the um our biggest users the the most successful ones are the ones where they've been very proactive and they've had people driving um both the training but also the comms and the messaging and it's had support from the top down it does need a lot of effort in order to make those things to make those things successful and and and actually people underestimate the uh um uh the amount of effort required and and you probably know this already but um Iran and I go to the Gartner conference uh which we jokingly called the Glastonbury of AI and um uh and when they started going on this year about change man it's all about change management we were a bit disappointed it's like well we kind of knew that it's so it's always been about change management or project management as Kira would say because he he thinks change management is just better better uh bigger project management so uh so yeah that that is a real that is a real challenge getting that getting that piece right and it does need um it does need both uh top down but also a lot of training and effort to make people to make people change but if they can't find if you can't find a way to to inject ai sensibly into their workflow they'll just carry on with doing whatever they've always done that's that's that's that's the case.

SPEAKER_00 34:58

I had another sort of question for sorry were you gonna say this McKieron I began my interruption no so I was going to if you carry on I've got I was I was I was gonna say Cam if we do this if we do this uh conversation in a in a year's time what predictions uh or what what changes do you think uh you'll have seen in in a year in in the work that you're doing I think uh that's a really good question I would say that um the workforce uh the teachers and and the faculty would be using AI in a far more considered ethical um and uh value added way so I think I think I think the environment would be more um would help them to use tools in a way that will enhance learning and support students in a in a in a in a positive way because I think the big the gap is with tools are all around us in education uh but there is very little evidence um or perhaps um some emerging evidence but it's limited on impact on student on student learning and I think that's where we want to get to I think if we're able to demonstrate that there is a positive impact on student learning I think that would be a huge game changer and I think uh I think we're shifting away I've noticed a shift in the last maybe six months I think the emphasis was all about tools you know all the all the new tools were out and it's you know all you know all the new tools and and it was all very sexy but now reality has hit home and I think we are hearing more conversations about some of the more practical if a bit boring topics about you know uh the financing the capacity building the policies the strategy and I think um I'm I'm I'm I feel quite positive because I think these are the right conversations to be having rather than let's just um you know uh blow out blow our budgets on on some tools that are being sold to us um and I think being more considered I'm really pushing for institutions to take a pragmatic pragmatic approach by testing getting those user cases then scaling them up um pushing pushing more for you know greater sovereignty uh particularly around around data and around tools so I think I think these are the I think the conversation will be more around um impact yeah and I completely agree with you I think that is the bit that's missing still and then partly education the big problem as we all know always is that it takes a long time to measure the impact of any intervention in education because how long do you give people to in teach in teaching classrooms you've got to train the teachers then the teachers deliver it and then you've got to give it some time to soak so you're probably a year or two after the fact before you can really see whether it's made a difference but I think that is absolutely the argument I hope we're gonna see uh teacher retention in improve I mean that tends to happen with economics as well so it's like most most social sciences it's not as straightforward uh the AI did the difference but um if we can see some of that that would be amazing as I would say the the 50% of teacher time that they spend doing admin that they don't want to do is the number one reason they leave and AI can handle a whole bunch of that and if we can do that and keep the best teachers teaching and doing the stuff they came to because they love um then that is true of social workers it's true of housing officers it's true of uh teachers and I'd love to do more of it absolutely I think I think just just on that just the last point on that you know I think I think also doubling down on on teacher quality is going to be a key part of this because there is this kind of notion oh well perhaps we don't need teachers anymore because AI can do all the tutoring and you can have your own AI um assistant you know helping you but what we've seen there's a really interesting uh research piece just come out from Stanford which shows or magnifies the impact of having a teacher as part of this process right because if you leave a loose leave a student just with an AI tool um and measure that against a student working with an AI tool and a teacher the the level of learning uh and level of attainment is far higher with a teacher involved in the whole process and I think that's you know and I do think that teachers role and I described it in the past is is is actually more of a learning scientist now than it than ever before. You've got to understand data you've got to understand the technology you've got to understand um behavioral science um it it's really really um you know far more um complex and nuanced than it's ever been before and I and I think the future of teaching is actually that of a learning scientist.

SPEAKER_01 39:57

Nice that's really good.

SPEAKER_00 39:59

I love it um we should uh we that's a great the learning scientist as teachers there's a there's a title I've already written it down we're stealing that we're having that camp thanks very much that was the first admission so my final question Cameron is um uh who's gonna win the football world cup so we are on to a semi-final today as we record this this this podcast is going to go out after so after after we'll know the result so um over to you for your predictions well uh my prediction is that um England will win the World Cup beating Spain uh on penalties I think it's gonna go to the penalties and we'll but we'll win the world cup beating Spain on penalties and so um that's that's my I mean I think I think we're getting to a point now that it's now or never I mean my ambition is to see England win the World Cup in our lifetime and with we none of us are getting any younger Kira so you we've seen England win the rugby world cup which I know is important which is you know which is really you know the sport that you you you follow in particular um and we've seen that happen but the World Cup in football is is is really you know um the dream yeah it would be nice yeah it would be nice and well I hope that you're correct I hope that you're correct me too prediction me too and that that would lead to that would lead to a bank holiday as well so I think I think it's about yeah yeah yeah well I'll be going down the bank with my carrier bags and saying Cam says you should fill these up for me so uh cool cam we've asked you loads of questions have you got any questions for us before we part our company for today have you got anything on on your mind that we can help with from your end well I mean look it'd be great to know about some of the great that work that you're doing in particularly the um tertiary education sector I know you do a lot of work with FE and I think I think FE is an incredibly underserved part of the education system. I think I I do you know one of the interesting things about AI is that you know the reports around how many graduate jobs that will replace is actually quite interesting because I think try and replace a welder or uh you know a w or you know um an electrician I think it's harder to do and I think that vocational education has got a huge role to play. So what are your reflections in terms of the work that you're doing so far with FE colleges and universities you know what are the key learnings that you can share with me?

SPEAKER_01 42:38

Yeah good stuff I mean I think the the I guess at a at a high level try trying to build the tools that really make an impact. So there's a kind of we do a bunch of hackathons with um FE colleges where that's where we'll get a bunch of teams and and do you know work through where their biggest challenges are and and sort of pick the things that AI can help with and then build the tools for them sometimes there and then one of the challenges is picking the right things because one one person or even one team's frustration may not be the thing that turns the dial and for colleges particularly now obviously learner learner outcomes number one thing they care about and that means achieving and uh enrichment activities um and employment kind of opportunities so that learners leave like that then alongside that is um winning more learners and you know that kind of recruitment making all of that part easier which is a place where AI can really play a big part because the whole enrollment process in that world is really challenging. So I think the the kind of at the highest level it's picking the right things to work on. The things we are working on is um the regulated Ofsted style you know Ofsted can inspect a college here anytime as you know and the colleges care about that stuff and in theory you'd hope that That's what Ofsted are looking for is the same stuff that will make learner outcomes amazing, you hope. It's probably about as good as anything you've got to go on. So we've got tools which help to uh diagnose challenges, uh identify missing data so that you can go and gather that ahead of time, and importantly, give suggested strategies for the things that you're struggling with using a whole bunch of um best practices from Ofsted reports from the EEF. Uh, we help create what the AI does. So signals here are two or three known strategies that will help fix that particular challenge for you to go and think about. So um, yeah, those are the kind of those are the big areas. The latest thing we're working on right now is timetabling, which in in tertiary and university is a really complicated picture. I'm told schools is relatively straightforward because I think because all your teachers are there all the time and all your students are there all the time, and therefore it's really about which classroom are they in uh and what order you do you do it in. Whereas for colleges, you've got session and unis, you've got sessional staff, you've got students that will be part-time that will be coming and going. So there's a and then room more specialist rooms generally, you know, laboratories and hair hairdressing and public services and healthcare and all of these kind of rooms that you've got to try and match as well. So that's um a really interesting challenge and taking us into a space which I'm quite excited about. We talked about it last week on the podcast briefly, is LLMs with more traditional algorithmic kind of machine learning style algorithm work. So you effectively get the LLM to help you work out what is the algorithm you need to be running, and there's some stuff then that helps run it, and then the LLM kind of picks up and checks the rules of be all been met and then does that. So I've kind of rambled around that answer, but I think the key is what's going to make the difference. Um, and yeah, in colleges, I think it is about more quickly identifying what's going wrong and suggesting intervention strategies, or indeed just kind of encouraging the intervention strategies you're already planning and you know giving you the data to show you where they're working, where they're not, etc. So it's quality improvement, really.

SPEAKER_00 46:14

And do you do you do you do you see yourself doing a uh a lot more work in uh curriculum? Because one of the challenges that institutions always face is how do they update curriculum with labor market intelligence, with all the other information that sits around? And I and I do think AI has got a massive massive role to play in terms of it. Yeah.

SPEAKER_01 46:35

So for universities, we've got some tools and colleges have it available to them as well on um labour market information and government skill strategies, and they've got federation of small business kind of stuff and employer fee, all kinds of um what other skills and they can be by an area or or nationally or whatever. Um, so that part, and we've um we've built with one of the universities a validation assistant, which is a way of you saying, right, here's my current program, and I need to put it through re-validation. Um, and that involves looking at all the skills and checking those. It includes looking at subject benchmark statements and ensuring we're meeting those, and in obviously then updating wherever's needed. So we've got AI that that does exactly that, it knows their validation process, so it knows what what they want us to cover, but it also has it's trained on our retrieval index, is all of those is 104 different documents. In fact, in that particular one, so it can go and find instantly the subject benchmark statements for marketing with whatever, and it can tell you what the labor market information says about the skills requirements and what the government strategy is so that you can build all that in immediately. Um, the aim being that's all you know, a lot of that's admin. Um in my well, it's it's you want the the program lead in that case, you want them thinking about how do I innovate in this and how do I make it exciting and how do I bring in, you know, which teachers or tutors, lecturers can I be using here. And you don't want them necessarily reading through loads of subject benchmark statements to be able to tick some boxes to say, yeah, it does that and it does this. So we'll let the AI do that and free yourself for the interesting work I would argue about you know, thinking about that course design overall to make it exciting for people.

SPEAKER_00 48:19

Fantastic. Yeah, really, really, really interesting.

SPEAKER_03 48:23

Cool. Well, we are we are gentlemen, we're coming to the end of our time together. Um uh I will leave you on a couple of facts. So, Kieran, I think you are a very big apology, not just to Mr. Mercer, but also to University World News. Oh no, which attracts over two million annual unique visitors and has 50,000 subscribers to its weekly global e-paper, according to Google. So I think it's a much more uh august um uh journalistic enterprise than than this one. So uh yeah, we we should definitely give them a plug, see if we can get our see if we get our numbers up, Kieran.

SPEAKER_00 49:04

You can expect a lawsuit from the University World News, Kieran.

SPEAKER_03 49:08

To be honest, Kim, they're not the only ones we're expecting lawsuits from. Maybe we'll edit that out there, mate.

SPEAKER_00 49:14

We can edit out of that piece. They do a weekly um like release and they they did it the other week. Um and the incredible thing, in the same article, they talked about my article and they did an article on the Pope, and it was in the same piece.

SPEAKER_02 49:32

Yeah.

SPEAKER_00 49:33

Well they maybe this is your future.

SPEAKER_03 49:36

Who knows? Who knew that you had you would have so much so much influence these days, Cam? Go back if I go back 25 years when you were a young, good looking, still are good looking, still are young, young man making his way in the world. Uh you've risen to incredible heights. And I just wanted to say congratulations to you. Long may your rise continue. And um uh and I understand you you might actually be in London next week, so we should definitely try and do a proper bar thing and get together.

SPEAKER_00 50:09

So it'll be fantastic to see you, gents. It's always a pleasure. And um, can I just say that um knowing you having you having knowing you both for 25 years and the role that you've played in in my life is never forgotten, and I hope we'll be friends for another 25 years.

SPEAKER_03 50:26

I I will happily drink to that. Gentlemen, it's been emotional. Thanks very much. Have great days both, and I will speak to you both soon, I hope. Cheers, guys.