← Digital Disruption with Geoff Nielson
Digital Disruption with Geoff Nielson
AI is Breaking the Economy and Education. Here's How to Fix it
01:13:19
Loading summary
Transcript48 lines
- [00:02]
A
What we're trying to do right now is take 24th century technology and shoehorn it into an economy from the 21st century. Instead of adjusting it, we're choosing to accept it as a foregone conclusion. Whereas my argument is, well, why don't we use this moment to reevaluate things?
- [00:20]
B
This is a show about the future of tech and the future of work. I'm Jeff Nielsen, and today my guest is Michael Clark. He's a longtime AI and transformation leader, serving at organizations like J.P. morgan and MasterCard. This is a guy who has walked the walk and led tech and workforce transformations many times over. Michael thinks our economy, education systems, and companies are built completely wrong and a better fit for the 1800s than today, and that if we don't fix it now, we're heading for disaster. From AI to data to leadership. I want to ask him what specifically we need to fix, how we can fix it, and what needs to happen if we're going to get the future that we want. Let's find out. Michael, thanks so much for joining today. Really excited to have you. Maybe just to jump into things. You know, let's. Let's zoom out. And I'm curious, you know, what the future looks like from your perspective. Where do you see the world going? Where do you see the economy going? You know, what. What is the world of 2030 kind of shaping up to look like from your perspective?
- [01:24]
A
So it goes in two directions, is the way to think about this. I don't think there is a linear path. I think where we are right now is in a very interesting dilemma or choice. So if we stay where we are and the economy actually doesn't change the way we classify it, the way we teach children, the way we define work and the way that we live doesn't end really very well if we think about it, because the metrics and the classifications of the economy, by 2030, if we follow the AI path, we'd have a lot of people out of work, a lot of high productivity, but a lot of unemployment because our measurement systems don't cater for the world that we're moving into now. If you flip the script and say we redesign education so everybody has the opportunity to participate. We reframe AI as a form of collaborative intelligence, we reframe the way that we measure the economy today in a completely different way, and we use this moment to reclassify a lot of things. We could have an incredible society in 2030 because we wouldn't have mass unemployment. We'd have people who are Multimodal, with the abilities to collaborate with technologies rather than compete against them. So we are on a very unusual moment in history where we almost have two paths to take. And right now a lot of people are talking about path B, which is the doom and gloom or opportunity, whereas nobody's really facing the path A, which is the more complex, but it actually leads to more balance and the evolution of humanity.
- [03:11]
B
For the purposes of our conversation today, I'd like to spend a little bit more time, I think, on path A. And part of that is not only because it's a happier path, but because I think, you know, to your point, the conversation around path B has kind of been beaten to death at this point. You know, everyone is talking about, oh, you know, AI is going to take our jobs. As you said, productivity is up, but it's just, it's not necessarily, not necessarily the world that I think we want to inherit. And so you mentioned this idea that we need to do, you know, some reclassification, we need to make some conscious choices. You mentioned, you know, big areas like, you know, metrics like education. I'm curious, what are the big blocks that we need to move or what are the big things that we need to get right or shift if we're going to consciously move ourselves from world B to world A?
- [04:04]
A
Yeah, it's a great question. So let's start with when we first walk the Earth, right? In terms of how we all learn, that's like the obvious place to start. So from an industrial perspective, we've all been taught to get a job. We basically are taught to go to school, we memorize things, we consume knowledge, and then we pass a test and we follow formal measurements, and then we enter the workforce and survival of the fittest. But typically, with the right domain skills, you will get by, right? But AI has thrown that completely out the window because AI now owns knowledge. So then the question is then what does the human need to do to not keep pace with the machine, but partner with it? So part of it is teaching children abilities rather than actually getting a job. So it's really about how do we teach children Critical thinking, the ability to debate, the ability to reason, the ability to unlearn all of these fundamental foundations even before they touch technology, because it's proven that the frontal cortex doesn't develop until very late in children. From 07 onwards, there's a reason why children don't know what's real and what's not up until that point. So if we want a generation to be able to spot Anomalies in technology. To be able to challenge it and improve it, we need to give them abilities that allow them to adapt. So it starts there and the reevaluation of the education system. So part of my work was redesigning the school's curriculum in this way. So we basically give children the foundations to almost become multimodal. Because you and I both know in five years time we won't be talking about AI. We could be talking about quantum or whatever other technology that's about to disrupt us. But the point is, if we teach children in the way that I'm describing or in a way like this, it doesn't matter what the technology is because they will be able to adapt with it. So the first big building block that needs to be addressed is to reframe the way we teach children so they enter the world that is not the world that was. So they have the ability to design the world that could be. That's the first thing that has to be done. Because we are teaching children things that won't exist by time they grow as adults if we carry on as the way we are.
- [06:34]
B
You've used the word children a few times and it's interesting to me. I want to just push on it a little bit. You mentioned the age of seven. I think there's a reading of this where you say, oh, by the time, you know, by the time of the education system, people get to university or college, you know, get it, get an arts degree, learn critical thinking and reasoning, take a philosophy course. It sounds like you're talking about something fundamentally different. You're talking about actually embedding this from, you know, a young age from, you know, primary school. Is that right? Like, like what is the age range we're talking about and how would this sort of be taught out, you know, over the education years?
- [07:12]
A
Another great question. Right, so you do this in two ways. Well, it's like imagine like a matrix. So it's across life stages, obviously, because you don't teach children abilities at a young age that you would as the brain develops. So you have to align this to how the brain develops in line with a child's age. So for example, you wouldn't teach like a 2, 3 year old critical thinking, right? But you would allow them to have play and creativity and various other bits and pieces and they start to build early life experience which are fundamentally foundational. Now, how do you teach this in two different ways? So one is, obviously there are going to be new school subjects that don't exist today that children will need things like data literacy, AI literacy, maybe even understanding their purpose as children. The other piece of the puzzle is embedding abilities into existing curriculum. So, for example, you could apply abilities into history. So this is an example I use in the book and some papers is that. Let's take history. Today you're taught to memorize some linear facts, some historical figures, and then you will go away and pass a test, or you'll write some essays and you'll do some homework. If you flip the switch and say, look, I've told you everything about World War II. Go away and come back, what you would do differently now, you know these things if you were in that person's position, that's a completely different set of brain skills. I'm now asking you to use. I'm asking you now to be resourceful. I'm asking you to come back and present an argument. But the cool thing is I can still test the things on the rubrics that I still test today, which is grammar, facts, all that stuff. But I'm now testing another set of ability, which means these are things, by the way, that children don't have today because either technology is ingrained or removed those skills, or they just don't have them. So the part of the puzzle with the AI journey is teaching people abilities they've either forgotten through time they don't have, or they have, and they just need to keep strengthening them. So we all, I would argue, don't have really good data literacy skills. But some of us are really good at critical thinking, for example. Some of us need to strengthen that and so on. So those abilities grow in line with the child's development and brain development, which then says, well, how does the classroom change? Well, instead of the teacher standing in front telling you what to do, they become a facilitator. And by the time they're about 10 or 11, this is the point where you would probably bring AI into the classroom because the children have the foundational abilities fundamentally to be able to operate a machine and complement it and be able to improve the models and be able to improve the technology.
- [09:56]
B
It's a really interesting approach. And by the way, I'm a big advocate of it. I'm, you know, I absolutely immediately see the value of more critical thinking, more reasoning, more, you know, data literacy and, you know, information consumption literacy in general. So let's, let's keep walking down this path so we, you know, know, we make some changes to the education system. You mentioned the idea of not just necessarily having the education system exist to, you know, get a job. So let's say, you know, we, we still, you know, we've taken children through this stage, this, this system until they're in their early 20s or, you know, maybe a bit younger, maybe a bit older. Where to next on this journey that we need to get right.
- [10:40]
A
So there's something that underpins all of this, which we'll get to, which is the biggie that needs to change. So the second part is when I go to work, right? So those abilities that the child was taught now become competencies because now they become part of the working competency framework. Because, let's be honest, AI in the future is not going to be looking for job titles, it's going to be looking for certain abilities that it needs for someone to be able to interpret its outputs and somebody to improve it. So the workforce now needs to start transitioning. And so yes, you still will have job titles and domain skills, because I still would want to probably go to a doctor. I probably would want people to have those domain skills which people will specialize as always. But there are some fundamental abilities that those abilities that children were trained on and learned through time, even the ability to continuously learn, become part of the fabric of the future. I hate the term, but human resources function, right? Because the work model, and again, a topic that no one is really talking about, is a work environment where a machine is also classified as an employee because AI is transitioning into the, as far as I'm concerned, a classification as a worker, not software, which means in the work environment you're going to have a group of people that traditionally manage transactions and people is now going to have to manage capability and intelligence because once you teach somebody the ability, they now become capable. So really, the workforce now is becoming capability driven and continually teaching people the ability so they can move around the organization and become more fluid. And eventually, which is the bit that underpins all of this, is they will no longer be seen as a cost, they will be seen as a form of liquidity because those abilities and those people will become competitive advantage in a world of collaborative intelligence, which is the combination of a human and a machine.
- [12:42]
B
If you don't work in enterprise technology, you can skip ahead or take a quick braincation, but if you do, infotech is your secret weapon. Infotech helps IT teams get projects done faster, better, and at a lower cost. Whether you're rolling out new tech and improving processes or just looking for cost savings, Infotech provides unlimited access to practical tools and expert guidance. You need to execute at A fraction of the cost of traditional consulting, no matter the project. From AI strategy to cybersecurity to vendor negotiation, InfoTech has you covered. Check it out at the link below. And don't forget to like and subscribe. Can you unpack that last bit a little bit more? So I just want to understand you're talking about this being an asset, this being a form of liquidity. You're referring to the people side, you're referring to the AI side, or you're referring to, as you said, collaborative intelligence and the unique, I guess, system that encapsulates both.
- [13:39]
A
Yeah. So let's start with the people side. So I've never heard anybody talk about this. So this is something I figured out four years ago. I'm sure other people have mentioned it, but I've never seen much of it publicly. Most of the issues we see today around technology, privacy, data privacy, data regulation, people losing jobs, all of this is a value problem. Even the point of view of people's data being taken or people or some companies making money on data and people obviously not being comfortable with that, it's a value problem. In fact, all of the things that we see today is an accounting problem because none of these things appear on the balance sheet. Data doesn't exist anywhere on any financial statement in the world, and people are still classified as a cost and an expense and being kept the equivalent as the machine comparably. And the reason, by the way, data isn't on, data flows through the economy today, every day. You and I both know that it powers nearly everything that we do, but there's no way of recognizing or recording that value and distributing it. And we are basically valuing the economy and everything that flows through it based on a 250-year-old industrial model defined by Adam Smith. Because the accounting model doesn't measure anything really intangibly even yet the world is mostly intangible. We still measure things tangibly. So when people are losing their jobs constantly, there's a reason for it, because we still classify as an expense even though we look at the cv. So the irony is once you can start training people on ability, they can be valued as well as can be the data. And even AI eventually will be valued as an employee because those combinations will determine what a future value of a business will be in 2030. And the reason I use collaborative intelligence is because we keep using the word AI constantly. I actually believe with the right abilities people have, it becomes then a form of collaborative intelligence. Because AI augments the human ability to go further and the reason all these job losses are happening, quite frankly, is because we still, from a value perspective, we still value people as a cost financially rather than the value they bring in an actual asset into the economy and something that can be measured.
- [16:13]
B
So let's talk about that valuation problem for a minute. So in your world, Michael, is the solution to this to find a way to financially value people and data and incorporate them into our existing financial structure, or do we need to kind of rebuild our accounting principles from scratch?
- [16:36]
A
No, you said, oh my God, no. I think that's a large on the last. Good, good, good. That's not a hill I want to die on. But I always believe this was never a revolution. This is basically an evolution of the. Exactly what happened with Adam Smith's framework, right? 250 years ago, Adam Smith created, you know, the wealth of Nations. The people that followed him didn't rebuild it, they just added to it. They added the concept of wages, they added various other bits and pieces. The thing is, at a certain point in history, it never changed. So, you know, there's no such thing as the digital economy doesn't exist as a term because the foundations and classifications of an economy never changed. All that changed was the value exchange in the economy of how value was exchanged. So now it's digital rather than physical. But GDP never changed as a measurement, and the accounting rules and the accounting models didn't change that much. Maybe we started to add brand and things, but there's not very many places you can put it into an accounting balance sheet. So what we're saying is what we need to do now is take a step back and use AI as a catalyst for change. And to say, right, let's look at the way we teach, let's look at the way we learn, but let's also look at the way we record value that flows through our economy. And let's think of the way we distribute it, because if we don't, AI is going to is. That's all AI is doing right now is acting like a mirror. It's exposing fundamentals that we should have adjusted 20 years ago when we moved into the digital era, but we never did. And what made big tech and very others very smart is they spotted the value flowing through the economy that other people missed and they found a way of extracting it. Because this is really important as well. In industrial age, value extraction existed, but it was tied to production. So you had to produce something to extract the value. The world we live in right now, I can extract value with no production I just need to sign up to a website and then they will just extract loads of value from my data and various other bits and pieces. So for the first time in history, value extraction is occurring completely separate from a form of production. So everything that you and I know economically today, most of it doesn't really naturally fit in terms of how the modern world works. And ultimately we are leaving a ton of value on the table.
- [19:14]
B
So it sounds like, if I'm following that thread correctly, we're talking about the idea of recognizing. It feels to me like we talk about the digital economy, we talk about AI, we're ultimately talking about, I think, data here possibly also, you know, the, I guess AI as a. As a reasoning engine or, you know, some sort of data manipulation tool is the idea to just crack open, like to monetize that in a way where we crack open a line item on the firm's balance sheet or wherever, and we say we have this much data, or we have just finding some way to put a number on that. I hear us going back and forth a little bit between firm level and sort of macroeconomic GDP level. Is it one, is it both? Where do we start this?
- [20:10]
A
Everything rolls up to a macro level at the end of the day because every business balance sheet rolls up to a central bank and eventually rolls up to a nation's number, right? So it always interlinked. Regardless, the realistic thing we need to do here is take a step back and actually look at our broader economy and what do we need to adjust to factor in the disruptive nature of artificial intelligence. Because at the moment, it feels like we're taking the easy way out. And the easy way out is universal basic income. It's like, oh, we'll just give everybody a check and it'll be okay. And the reality is we've got the greatest minds, have created the greatest technology, but yet we don't have the intelligence to address our economy and to make the changes required. So we all flourish rather than flounder. Because what we're trying to do right now is almost take 24th century technology and shoehorn it into an economy from the 21st century. And instead of adjusting it, we're choosing to accept it as a foregone conclusion. Whereas my argument is, well, why don't we use this moment to reevaluate things? So to your point, let's just take data. Data can actually be valued. I've come up with ways to value it and so have many others. So now that becomes an asset class on the balance sheet. I can value people's abilities, their training courses, all the things that they've done and all the things that they bring, they shift from the expense column to the asset column. So now as a business, I actually know where the value is coming from in my business. Because let's take the example. If you left your job tomorrow, okay, maybe you're a bad example because people know what you do day to day. But most people who leave a job, all they, all the management understand is the cost that they bear. They don't know the relationships they bring to the customer. They don't know the other things that they do or the abilities that they've built up over time that are possibly competitive advantage for that business. So this is invisible value that just walks out the door that could take five years to come back. And AI is exposing all of this because we're falling into the trap of looking at this as a cost and efficiency play to reduce workforce, because AI makes things quicker, but as a consequence of the things that AI needs to improve, which is effectively the abilities that people bring. So this is a broader. Let's take a step back, let's look at the bigger picture and let's actually rethink the economic measurements and the things that we value most. Because I think that is not in anybody's conversation right now.
- [22:57]
B
It's, it's really interesting to me and I find myself coming back to the comment you made earlier about the fact that in our modern economy, so much of the value being created in the economy and certainly the value that, that the workforce is creating is so indirectly tied to any sort of obvious production. Right. Like we're, when we're talking about the knowledge work economy as a bucket, it's not people, you know, building Model Ts, it's not an assembly line. And we've just, you know, I think a lot of leaders, if you press them, would admit they would have a hell of a time actually mapping that value or understanding that value. It's so nebulous right now, like, oh, what do all these people do? And to your point? Yeah. Well, we don't have a good answer, but we do have a better answer for the cost. Right? That's what we know. We don't exactly know what these people do, but we know how much they cost. So with that sort of problem and that framing, how do you envision you get to the other side of the ledger in a way that I guess feels intellectually honest and defensible? Because we're starting from a place where we don't necessarily understand the value. Now, I think it's pretty obvious that there is value there. But how do you start to build toward a world where I guess before you even get to a number, you understand the value being created?
- [24:24]
A
Yeah, it's such a big question, right? I've been working with the International Valuation Standards Board personally to start. Because at the end of the day, without a value, you can't derive a price. Physically impossible. So where do you start? Well, you start with the people who actually define value for the industry and you, you basically. And then you have to define policy, then you have to design principles and et cetera. But I actually don't even like the word policy anymore because the world's moving too quickly. We almost want frameworks, right, that can adapt. Like the way you can start is you, like, like anything, you start micro. You have to. Can you start valuing bits and pieces of what people do, like the abilities they bring? Like the end of the day, everybody gives in their resume, like what, what are the abilities that they're bringing and how do they fit into the competency model? And how can you value. Can you start to value some of those? But the truth is, look, I can do all of this in isolation. None of this happens unless the right parties come to the table in the party like you. You've got to start engaging with the accounting boards at the higher level to see how we can start adjusting the balance sheet. Because at the end of the day, all roads lead you there. Our whole world, whether we like it or not, is run on a balance sheet. Like all the layoffs we see with AI. Yeah, you'll make some savings, but in 12 months time you'll be paying more tax. Because there's a. I don't think people understand the second and third order effects of some of these decisions. So I lay off 5,000 people because of AI. Those 5,000 people pay tax. National debt is really high in some countries, interest rates are so high on the debt that they have to borrow more money as a government to pay back the interest. I now have 5,000 less people paying tax who are out of work. So now I may have to increase forms of tax on those businesses to cover the cost. So whatever you save now in a year's time, you're now paying back because of the circular nature of how all of this is connected. So everything has to start with the state. If you want to talk accounting terms and you have to agree some fundamentals and also some proof points because you don't do this big Bang, we're talking something. This is huge, right? The conversation you and I are having, we're effectively saying we need to reclassify pillars of an economy that are decades and decades old. And you don't do that with a light switch. You pick it piece by piece. But at the very outset, you've got to have agreement as a collective, whether that's an AI consortium, whether that's a group of leaders from all elements, not just technology, that fundamentally take this on and test it in areas, because it's the only way.
- [27:16]
B
So is that step one? I wanted to ask you, where we start with this initiative. Is step one actually deliberately putting together a body or a consortium that can start recommending standards?
- [27:28]
A
I think it's more than that. So I wrote about this recently as almost like this. Look, what's happening right now is if you look at government now, I can't speak for the rest of the world, but I'll pick on the UK and the us. There are various departments that focus on different elements of a citizen's life, which is normal, right? They're all looking at different elements and also other people looking at foreign policy. So everybody has their own agenda and their own lens. And AI seems to be bumbling on in the background as everybody else is trying to sort out running a country. So they clearly don't, can't give 100% of their attention to the issues of AI today and what needs to change. So it's almost like there needs to be another independent, more adaptable body whose sole existence every day is to look at AI. But look at AI not through regulation, but look at it through markets, look at it through technology, look at it through the economy and then push it out. But it's a consortium of everybody that needs to be involved as a coordination layer almost, because we don't have this today. We're either trying to regulate it or we're just discussing it. We've got no one looking at the first order, second order, third order effects and the macro picture, but then also then being able to go deep in other areas and work with other government entities to implement and oversee these things.
- [28:54]
B
I want to come back to the distribution problem of AI, you know, using the economic lens that you mentioned and where that comes into play. And, you know, I think people talk colloquially about, you know, there's basically five big AI companies, and even that I think is generous because I think there's probably two or three that make up, you know, kind of the lion's share of the conversation and the value and you know, to me, this has really profound implications when we talk about value flow and we talk about capital, whether it's the fact that they're all American and any of the kind of digital or data sovereignty issues around that, when it's just concentration of wealth in these companies. And to me, if we go down this model, Michael, of valuing these things, I expect that if we change nothing else about the distribution, we're going to have a holy shit moment where we realize how kind of, you know, how tied or enslaved we are to the companies actually creating these tools. So I'm, I'm curious from the distribution angle, what you see and if you see a need for change there as we kind of move on to track
- [30:08]
A
A. Yeah, 100%, it has to change for the reasons you've described. And look, there are ways of data value to flow freely is called licensing, and it's called digital licensing and tokenization of data. We have all the tech in the world to do this today. I, I could literally create a smart contract because the other thing that people don't really understand, and this has come up so often in the last 20, 30 years, this whole conversation of data ownership and, and the reality is no one owns data whatsoever. You can't, you can only own the insight that you generate from it because it takes two people, sometimes three people, sometimes four people to generate a piece of data. So, for example, I'll give you a real world example. Take your banking transactions as an example. Well, I can't generate those transactions without a credit or a debit card from a bank. And the bank can't get those transactions that I generate unless I use the card. So in that example, neither of us own the data because we can't live, the data can't exist without either of us. But what I do have the right to do is have the access to those transactions to then generate my own insight from them. And I might choose that insight for myself, or I might license it for other people to use and I might make some money from it. And maybe I have an agreement with the bank that we make. I give them a percentage when I sell it. So. And a license then allows that freedom of movement and the rights around some of that content. This is the bit that we're missing in all of this. The whole fear story about AI taking over humanity. And the reality is the real engine of all of this is the data itself, but we don't treat it as a national asset. It's bizarre that our culture and our history is now recorded digitally in data, but yet we recycle it constantly and we don't treat it as an actual asset. Because, look, when technology fails, in my view, that's not a tech problem, that's a data problem, because people can't get the information they need to either for their Google Maps, they can't get it from signage, they can't get it from where they need. So my concern, when you talk about centralization control is not so much the tools, it's almost this data is free flowing into four or five places and then being controlled in those four and five places. And more worryingly, what comes out as an interpretation is also driven by those four or five places. And a lot of the conversations I've had, and I helped write the regulation in Dubai, for data to become a real world asset to at least, at least start the conversation, is that countries need to look at not the technology as its assets, but in a way it's people in his data because they're the representation of that nation. And if you start looking at it that way, there's also a big change here. Right, Ask yourself a question. Your utility provider is classed as national infrastructure because it's deemed to be critical to the running of a nation. It's the water company, the electricity company, they can't leave when they want. So the question is, when does AI software, cloud computing become critical service infrastructure? Because I was critical to an economy, which means then it comes into a whole other set of regulations. So the bigger danger in your comment is we're still treating these companies as software providers, not as providers of national infrastructure, which is very, very different.
- [33:44]
B
I agree with you. And one of the ideas that I've been toying with, and I'm curious for your perspective on it, is that the idea that at some point, AI or data or this kind of compute infrastructure, at some point, it feels inevitable to me that it's going to have to be treated like a utility and we're going to have to view it in that lens. Do you agree with that statement?
- [34:09]
A
One million percent. I can't believe it hasn't happened already. Like cloud computing. I don't want to sound controversial, but it feels like unregulated banking at times because there's no regulation in terms of what goes in. Like if I lost a thousand pounds in my bank, they have to give it back to me. That's kind of the regulation. And they've also governed around what they do with it. There's nothing bound in any of that data that goes in. If they lose it. I'm sorry. And I can't control what they can do with my metadata. But to your point, they're serving nearly 90% of businesses in our economies. It like that's more prevalent than some water companies. So it almost by default is a utility, but we refuse to classify it for whatever reason, because I think we still. Governments bless them, I don't think still understand it in the way that you and I are discussing. And I still think they believe they're just a software company that just provides software. But they're not because, okay, you could get away with it with cloud computing and SaaS solutions. When you've got a piece of technology which is disrupting every single industry more so than the Internet, and in some cases is looking and smelling like a worker, that changes the conversation radically.
- [35:34]
B
There's a whole suite of implications there, from, you know, regulation to taxation that, that I'm sure we could get into and you've probably given some thought to, but I may. Maybe where I'd like to go next is if you look inside almost any organization in my experience, and you look at the way they treat data, it becomes immediately obvious that they haven't. That they're not really doing any sort of internal valuation. Right. If you look at the way they treat, you know, it, or, you know, whatever you want to call the custodians or, you know, you know, governors or managers of any of the data, you know, their ability to understand where value lives within data, to generate insight. The entire system is, you know, to use your word, unregulated and, you know, just really not set up for any sort of value extraction or as though value actually lives there, there. I have to believe that there are organizations that are doing that better and have sort of figured it out, even if we haven't created these superstructures. So I'm curious, Michael, in a world where we don't have these governing bodies and this proper valuation yet, if you're one of the enlightened few who's listening to this and say, yes, damn it, I have this data, I know it's valuable, what can companies, organizations, governments be doing right now to better, you know, treat data as an asset?
- [37:06]
A
So as a. This is like, now you're going down a really big rabbit hole, but it's a good one to go down. So let's just start with the Chief Data officer, or as I like to call them, the Chief Governance officer, because really all they do is protect data, right? They don't do anything with it. They don't innovate with it. It's just tracked as a risk. A lot of boards actually don't even have data people on their board. They don't even recognize it as an asset. So for me, the first thing any business can do, if you have somebody leading data, they need to go on your board because that data is your differentiator. It's your competitive edge of everybody else because it's the custom information your customer gives you every day. Like without that information you haven't got a business. So it's make or break who you become today and tomorrow. It's why your, your competitors fight for it every single day to get a piece of it. Because they want the data that your customers give them. Right. Vice versa. So that's the first thing. The second thing is look the way you value data. And I had to, this has taken years and advising companies on how to do this and working with valuation boards and things. Data's a bit weird because it's not like any other asset. It's a bit like Ethereum and Bitcoin combined because it's a store of value and it's also utility. They can be both. So first of all, data has intrinsic value, which is the value internally to the business. If it's you as an individual, it could be personal value, it has extrinsic value, which is demand based. But it also could be culturally significant. It could be unique. You know, all the qualities ironically existed before the industrial age, right? Because we're getting down into the weeds of bartering traditional value metrics that existed before mass production. Because all of these are very relevant to data as well as whether it's portable, whether it's interoperable, whether it has high quality. So all of these things can be measured in a business today. But you wouldn't just go in your server and just try and rip everything out and value it because you'd be there until tomorrow. With some companies, no, you just pick slots and you extract your data, you structure it, you value it. Frameworks that you can use and you start slowly. But the first, honestly the first place a business should start and I've helped, I've advised hundreds of thousands of companies to do this is at a board level. You recognize it in the very instance as an asset to this company and as part of the business strategy. It doesn't belong to the technology strategy. Data should never belong to technology. Technology is custodians of it. They just store it and they just manage it on behalf of the business. Has to be part of the business strategy. It has to be part of the board and then it cascades down from there. Look, like you, I've worked for companies where I would go in as an advisor to the board. I've seen companies with like 60 years worth of data. I haven't got a clue what it is, but the only safe option is to pay the big tech companies to buy more storage. And by the way, if we value data properly, we wouldn't need as much compute as we have today and we wouldn't use as much energy and we wouldn't need as much storage because we'd only keep the data that was valuable. This problem has proliferated since the 70s and it's just got worse and worse and worse. And all we've done before AI is we built software to fix software. We never actually fixed the data problem. And ironically, data is the most valuable asset in our entire economy, proven to be up until about a trillion dollars by 2030, yet which constantly told is not valuable and yet people go to court for it. It's ironic, isn't it?
- [40:53]
B
It's really interesting to me and I'm right there with you on the value here. One of the issues I find, I guess with the board is when they think about data as an asset, they don't really understand it well enough to even know where to start, how to start thinking about it, where is it generating value, how can we start to, you know, capture this?
- [41:19]
A
Ironically, you never start with data because it's like going swimming in an ocean and you'll never get anything. Most businesses, I ask them, like what question are you trying to answer? And let's find the data to answer the question. And then you start answering a question at a time and you start valuing it the other route, which is a bit more. It's not messy, but it can be done. There are valuation frameworks that I can help a company pull together where you can start evaluating it. Part of the work I did with Evaluation Standards Board was, I think where you were going is you value data based on structures and categories. So you have raw data bits almost, which is basically the lowest form of data, which is just a mess. And then you have data sets, data suites. You have data sets which are intelligence ready, which means those data clusters are ready for artificial intelligence. And then you have others then, which then get higher and higher in terms of knowledge sets, which is really an LLM. And then you go even higher, which is then federated data sets. So give you an example, if I had two businesses next door to each other. One business had, I don't know, 20,000 data bits. So raw data. Business B had 20 intelligence ready data suites. That business is worth way more because it's ready for, for AI right now. That other business is going to be there forever. They don't even know what's valuable. They haven't got instruction properly, they don't even know.
- [42:49]
B
So, so that's, that, that, that's exactly where I wanted to go and to talk about data quality and the fact that so much data in so many organizations is not AI ready, it's not insight ready, it's just, it's a mess, for lack of a better term. And one of the, one of the cardinal sins that I've seen on boards, or at least of leadership that's not, you know, data literate is oh well, can't AI fix that for us? Like can't we just, why would we fix that ourselves? Because aren't we just three or six months away from AI actually you know, fixing this for us? And so I wanted, I, I wanted to, you know, pressure test that hypothesis with you and then assuming you don't agree with it or don't agree with it entirely, what is the better approach right now for, for getting your house in order? Order data.
- [43:43]
A
So I have to be fair. There is one company that is using AI for data discovery. So there are, there are companies doing this right now, but then that's only to help data scientists find the right data sets that they need within existing data sets. That's not necessarily value driven, that's more utility based and use case based. Right now when I usually advise boards on what I was doing in my old life, I still do today, I use a model that's used by Microsoft, it's used by mit, which is what you call a data race. And you literally do it one question at a time. So from a board perspective you'd say, right, we need to find a question we need to answer from data to prove its value. Because in the day, I remember once one of the biggest questions I got asked at an event was we can't justify the ROI in a data project, so we never do them. It's hilarious, right? They go home with Google Maps, but yet they can't do an ROI for a data project. So I said, look, the best thing for you to do is build your data muscle one question at a time. So you basically find the question, you define the measurements that you want for the question and then you start the race. So then you go, okay, so what capability do I need to answer this question and do I have it? And you might say, well, okay, I need to maybe, you know, partner with someone or whatever. The next important question, probably the most important question before the capability is what data do I actually have to answer the question Question. And you might be able to find a segment of it. You might have to get some in to then amalgamate and aggregate to that data set. And then you run your insight on the AI and you use it to feed your model, you train it, you create your inference and off you go. And then, and then you generate the output and then you finish the race. And then you look at what you the measurements were. Did I succeed? I have now a working model and now I can answer another set of questions, but I also have a baseline set of capability. Because the biggest danger with this, the AI conversation today, is we look at it like CRM was in the 90s. We look at it, what workflow was in the 2000s. This is the silver bullet to all our data problems that we've never addressed. Because AI is going to figure it out. When the reality is garbage in, garbage out, right? Data, the AI is only going to run in what you give it, and in the end it's going to hallucinate like crazy. And if the model is too big and it's not controlled enough, and not many people, by the way, spend enough time on inference data, because I don't think they understand how that needs to be structured in a way so that you can learn and improve the model. So there's a whole minefield that, that needs to be understood. But it starts very simply. Like, I did a presentation to a very big bank in Eastern Europe and they said to me, look, because of your conversation, we put data as part of our board strategy. And then all the doors opened after that because then it became a board conversation, not a technology conversation. And as I always say to them, look, you can always go all in on data at a board level, but you would never go all in on an investment level. You would do it incrementally to prove your value. And that is always my advice is you incrementally do this with a clear defined scope and a clear set of measurements, and then you will start to identify value and what is waste. I think the important thing after that though, is learning what to keep and what to destroy once you understand value. And that can be difficult.
- [47:27]
B
So just following that, you know that road for a minute. I'm sure you know I'm throwing a lot of stones here about organizations who are not doing this properly. I imagine in your travels you've come across at least a few organizations who have figured this out and are doing it better. And maybe not across every metric, but in some capability for the ones that are farther along this journey. Where does it land them? What can they start to do with this if they've actually been able to harness this power? Either things that are, I guess, more obvious or less obvious, that are capabilities that are emergent from this approach.
- [48:06]
A
So I can name one if you like, if you want me to name a company. So the favorite one I use all the time is Schneider Electric. So look, there's three ways you can generate value from data. It hasn't changed for decades. The first way is you obviously use it to reduce cost and improve. So you use the data to guide you of where the parts of the business that are underperforming and you make those improvements. And there are examples of companies who've saved millions of dollars by following the data to advise them. The second way is using data on top of existing products to give them more stickiness, to give them more insight, analytics, all that stuff, which then starts to generate more revenue. So one side cost reduction. The other is revenue generation. The third is the most controversial, but Schneider do it very well, which is setting up an entirely new revenue line of data monetization, but anonymizing that data to do so. So it has no customer information in there at all. It's just data that's grouped into themes. So what Schneider do which and they make money in two different ways, which is very clever. There's a few companies in the US that do this too. They take all the data from the machines and they're able to turn that data into APIs and they sell it on a data marketplace or like a data exchange. So developers can come in, they can look at data by theme, so it could be by weather. They can combine data sets, they can drill through graphically, or they can export that as a CSV or as an Excel, or they can actually export it as an API and put it into their apps themselves. So every time that API is called, they make money. Schneider. And the cool bit then is they almost did what Google Play does. Say the developer develops something that Schneider like, they can publish it back to the app store. So I'm making money now on the data that's being generated from my machines by developers consuming it and making money by the app that comes onto the store. And I'm also expanding the reach and ecosystem of Schneider Electric. So now I have a completely separate business. Now, if you want to name a company now, what you can do with data, right? That's just a very simple example. The most odd example that you would never think of is John deere tractors, right? 20 years ago, sensors in the ground, autonomous driving tractors, have more data about agriculture than the entire US Government. You imagine you have all of our structured data. And a few years ago, they were publicizing their AI journey so they could build an entire intelligence layer on top of all of their data to give real time orchestration and feedback to all of their farmers and generate an entire new economy on top of all of this structured data. Because this is where all this leads. AI with no data is. Is like you and me with no English language. But if it. If it's abundant and you control that moat, my goodness, then the fun really starts to begin and propositions that probably you and I have never thought of will be created.
- [51:22]
B
You described that last scenario as controversial, and I understand why. And I agree with you. And I also find it very, you know, compelling. And I mean, the other word that comes to mind, I think for me, and maybe for a lot of leaders, is seductive in some way, because there's. I think some leaders might worry that it's taking them down an evil path, if I can very deliberately use a provocative word there. And for me, the controversial part of it is the privacy piece. And, you know, let's step outside of Schneider Electric for a minute. If you're a bank and suddenly you're selling customer data and you can swear up and down, it's anonymized. It's very easy to imagine a world where people say, whoa, why would I want to do business with you when I'm paying you and I'm the product, you're productizing my data, is that this is probably too big a question, but is that surmountable for all businesses? For some businesses, does it mean this for most businesses is a door that we just shouldn't open? If you're leading an organization and this is an idea that comes into your mind, how should you go about answering that question?
- [52:39]
A
It's difficult. Like, I've had pushbacks from different organizations, particularly in financial services. For that reason, it normally falls foul of the regulator. The regulators, actually, I say it falls foul. They get nervous. But if you can prove that it can be securitized and it's private, people tend to be okay. But it's also how it's presented. Look, there is also a fundamental gap today in the model I've described and it's something I had to resolve in the writing as well, which is there's no rating model for data, is there? And there's, and there's no certification model for data. There's no watermarking to say that the machine that generated this, the data came from this location and it's rated to be used under these conditions. And this human impact is in consideration. So for example, let's just take our data exchange example. Let's just say we had rating. Let's just say I had a data set that was aaa, which means I know where it came from. It was generated in real time by a human in conjunction with software. It's been certified as validated data. But from a human impact it contains information that could identify someone. So the advice from the rating would be you can only use this information for internal consumption only. We don't have those guide rails on data today. And again, it's a big black hole because where you're going is that a lot of these decisions around the options I'm giving is based on the temperature check within the organization and the type of organization and industry that they're in. Some people do this incredibly well because they own the industry, like Schneider own their ecosystem because they're not selling data from without any in, they're not bringing data in. They do it within their own space and they know they control that space and they. But it's not financial information, so it's a bit different. It's information about equipment which is kind of rudimentary. Like once you start getting into spend information and things. Look, BBB had done a great job of this, it works really well. But a lot of the banks I've been to are very nervous for obvious reasons. But again, once you start treating data as an asset, can I value it, can I rate it, can I certify it? All of a sudden things start getting a bit more controlled. So whilst monetization can happen in a controlled way, again this comes back to everything, comes back to those foundational pillars of jobs to be done to actually make this stuff be able to flow freely, securely and be trusted. Because that's the keyword trust.
- [55:28]
B
Let's shift gears a little bit back to that system of capabilities and how we do this within an organization. And you've talked a lot about workforce transformation, about leadership skills and I have to imagine in your work you probably want to start with some of these fundamental questions about where value lives. But you know, can you share a Little bit more, Michael, about when you think about workforce transformation, when you think about AI adoption, when you think about leadership, what sort of the framework you use to help guide boards, where they should focus or what approach they should take.
- [56:05]
A
I don't think it's changed that much. I think the only thing I not really framework per se, but there's things fundamentally that leaders, I always use the line like clarity before consequences. They first of all need clarity around what they're capable of doing and what they have and what they don't have and what they can realistically do. And I would argue, why are you adopting AI? That seems a really stupid question, but a lot of people do it for FOMO or they want to do it because they want to appease shareholders or because every employee under the sun is knocking on the CEO's door to say we should be using AI like our competitors. But ironically, if you speak to a CEO, they will say to you, not all, but some. No one can tell me why. No one can tell me the impact I'm going to make by using AI on my customer, on the business, it's no, no, you just need to use AI. So my first question to them is always, well, why do you want it? Because don't forget, the answer is not always AI. It can be remove a process, it can be workflow, it can be whatever. It doesn't have to be a machine. That's the first thing. The second thing is incrementally rolling AI out. How are you going to do that? And for the board, I think they underestimate the organizational change that's on the horizon. So there's an educational piece there and of course over communicating because at the end of the day, people fear change. And the net rhetoric around AI is they're all going to get fired. And the reality is it's down to management and leadership to guide. And let's just put leaders aside a second, because a big conversation with them is understanding second and third order impact of decisions. Because this is a failing. In modern day leadership, decisions are made based on what's in front of them. They don't understand the second and the third and the fourth order impact of their decisions. So there's a framework in terms of how the business needs to evolve and how you incrementally do that. But then there's a framework for how the board then makes its decisions on what AI is changing and what needs to be done and how they need to change as leaders in the first place. Because most leaders today are still treating this like software and underestimating the impact of change. I don't think, not to be disservice to the modern mba, it doesn't cater for the world. We're asking leaders to redesign. Right. Because most of them are being taught accounting, not transformational change at the magnitude that you and I are describing.
- [58:47]
B
The word that comes to mind for me, and I don't know if you would agree or choose a different one, is the cultural impact and the cultural dimension of it. And one of the biggest barriers in my experience is that the expected pace of work being done and of decisions is just getting faster and faster. And the faster that you're expected to make a decision, the less space there is to consider consequences. It's just we need to get something done, ship it, and we'll figure out once it's in production what those second and third order consequences are. Do you agree with that framing? And if so, how do you break out of that trap?
- [59:33]
A
I agree with it. I literally wrote about it a few months ago because it's what I've observed in every board meeting and every meeting as a former product guy as well. Like, everybody knocks on your door and says, we need to deliver this feature yesterday. And when you really push them, it didn't really need to be delivered yesterday. Look, as a leader, I think you need to give yourself space. And I always try to get this is me personally is like, I try to understand three or four critiques because leadership for me is about asking better questions.
- [60:06]
B
Yeah.
- [60:07]
A
So if someone came to you, so we got to do this now, what better questions can you ask to find out why I have to do it now? Like an old mentor told me, trust but verify constantly. Which is the first rule of leadership, I think, is trust but verify. If someone says, I've got to do this now, okay, cool, I trust you. But I need to know these three, four things. If I do this, what do I impact? And that's where your people around you make a difference. If I do make this change, what effect will I have later? Because every decision is about impact. Because I was taught that leadership is about impact. Whatever you do on a daily basis will have an impact, be it to your customer, the business, your shareholders, even your own personal reputation. So as a leader, I always believe that everything rolls back to trust but verify. And then asking a set of better questions around the thing that someone's coming to to make a decision. Look, the other problem, and you've used speed as an example. The second thing, which is the biggest blocker to innovation. Change is risk tolerance. Most leaders today are not the leaders of 20, 30 years ago in terms of what they're exposed to. If a leader makes a bad choice today, they're on the news, they're on social media, they are shouted out by shareholders. Their own personal reputation is destroyed in five minutes. So some of the big decisions that a company needs to make, many leaders are afraid to do so because they also, their tenure is so short. Most C suite leaders, 10 years, like three years maybe. So not many leaders are brave enough to take on something big because they'll never see it through to the end. So if you. One of the big things, if you want to roll all this back, is when it comes to leadership and rolling out some of the changes you and I have been discussing what needs to change for a leader's tenure and what are some of the responsibilities they need to either see this through or hand it to someone else. And does that mean we need to measure them differently? Because that's a whole other debate in itself around what a leader needs to be in what I call the intelligence age.
- [62:32]
B
You got my wheels turning there. And there's sort of an AI angle, but it feels like with the measurement system right now, what we're like, what we're architecting for is leaders who make the most average decision possible. Like, how can they do the least controversial thing and upset the fewest apple carts? And just the exact average decision. That's what they'll do because nobody can yell at them for doing that. And to me, it's like I'm chuckling because AI is literally like, what is the most plausible answer? What is kind of the average of all the information we've taken in? And. And it just feels like a huge mistake. It feels like something that is the enemy of innovation, the enemy of big bets and actually having any sort of transformational level change. So how do we architect away from that? How do we architect to a world where you mentioned the idea of leaders who stick around to see their change through? Is that the key piece? What are the key ingredients so that we can have, I guess, braver leaders. But to your point, also leaders who are going to do their homework and make sure that they're not just, you know, doing something that is, you know, brave but foolish because it's splashy.
- [63:48]
A
Yeah. So look, I think. I think there's a few strands there. The first thing is I think we need to revisit the way we teach them to be leaders like MBAs and things today, Bless them. And all the extracurricular stuff leaders do, they're based for a world of industry where we made widgets. They're not made for the world that we live in today, which is fully digital. A multi generational workforce which demands different levels of communication. Leaders need to be system thinkers today. They need to understand the connections between things to be able to understand the impact of what they make, which means they themselves need a whole set of abilities that they're not being taught. They don't, they don't need to be told the basics because that's just what it means to be a good leader. They need the fundamental skill sets to do it. That's the first thing. The second thing is leaders need a different measurement system than they have today. The way that a business measures a leader's success can't be just based on a simple set of KPIs, has to be done based on impact or something very different in terms of what success looks like. And then, and then thirdly, look, some businesses have put leaders into golden handcuffs deals and said, right, you're not leaving until this is done. Which is harsh in some cases because some change is never done, it just continually evolves. But some change can be time boxed. And some leaders are told to stay until that's finished. But also there's. The leader must see it. If they, if they can't see it through, then their impact must be based on what they did to hand it over to the next one. Because, and of course they're not forgetting how big salaries are with leaders. And I'm not suggesting we lower salaries because I'm not going to get into that political conversation. But I think there are some fundamental dimensions of what makes a modern leader in the. I wrote about this in my new book, which is the intelligence economy, because it is actually an economy. We need leaders that are ready for that, that are able to lead a multi generational workforce who demand a different level of communication by age group and are able to understand all these different headwinds and world and things hitting them on a daily basis and can be pure system thinkers and have the right people around them to then make those trust verifiable decisions with a whole new set of measurement systems. We basically need to teach leaders to be ready for the intelligence age, not the industrial age, which is kind of where we are.
- [66:20]
B
So to sort of bring the conversation full circle. You know, I'm curious in the intelligence age, which are the skills that are going to be done by the AI side of the kind of collaborative intelligence versus the human side, and you opened the conversation by talking about critical thinking and, you know, reasoning and, and some of those skills, are those the exact same skills? What are the most important ones that the leaders really need to be selected for?
- [66:52]
A
One big one for me is judgment. Because look, if you look at both sides of the coin, this is what this is, right? I think we need to finally accept that the machine now owns knowledge because it's consumed most of it. And the reality is it can also do pattern matching faster than I can and you can. And it basically has a fundamental set of core ability that we cannot compete with. And why should we bother? Because on the right hand side is what we consume as outputs from that pattern matching which we as leaders need to be able to determine. What we've been given is true. So judgment comes into play. The ability to then think critically around that judgment is massive. You'll be able to then take that and be able to debate it and use it and refine it. For me, those are the massive ones. Because ultimately this all boils down to decision making, the ability to make solid decisions based on machine driven output. Because in my mind, the output is the input into human decision making, an action. And if you can't evaluate that and you can't judge it effectively, how on earth is the model going to improve if you just accept everything that it gives you? So for me, if you want to be very philosophical, the human side is wisdom, the machine side is knowledge. That's really what this is, because wisdom is, is executable knowledge. So the machine now is generating all the knowledge that the human was taught to do 30, 40 years ago in a classroom and pass a test and spot patterns. The machine does that. Now the human needs to be able to now do what the human was never taught to do and has forgotten through time is to be able to actually turn knowledge into something that can be actioned and challenge it, use judgment, to use it wisely, and also have the ability to then improve it and verify it. Those are, those are the, those things combined get you to collaborative intelligence.
- [69:02]
B
I really like that framing. I like the framing of leadership as better decision making, how we can make better decisions, ask better questions. I like the idea of bringing human wisdom to the knowledge that machines bring us. As we start, Michael, to wrap up the conversation, you know, we've covered a lot of ground here today, from the education system to, you know, rethinking the economy and finding new ways to, you know, actually value data, to what's going on in organizations with their Data with their decision making, with their leaders. For any leaders listening to this, if you were to give them one sort of capstone piece of advice to take away from our conversation, what's kind of the most pressing thing that you'd want them to take away right now?
- [69:53]
A
I think that don't look at AI as a way of reducing your workforce. I think, honestly, I'm aghast with the amount of daily conversations and reports around people being made redundant. Job loss, job loss, job loss. I think we, we are underestimating the role that people play. And I think your people will be your competitive advantage with AI, not the AI. And I think the advice to leaders will be use this moment to upskill your workforce, to be able to use the technology and not use the technology to replace them. Because at the end of the day, we should be getting to a place where we're delivering intelligence with a human touch, because that's actually what the modern consumer will want at a certain point. So my cornerstone advice with them. Well, and actually there's probably so many things I could say to them. I think the, the most obvious one, other than just don't get rid of people completely, is understand why you're using artificial intelligence and these technologies in the first place. Don't just use it because someone told you to go into everything you do with clarity, because if you don't, there will be consequences, as always, because every bad decision leads to a consequence. Not now, but maybe two, three years down the line. So I think it's always go through these things with clarity. Don't implement ii without understanding why and the impact that you would make. Don't go all in. Do things incrementally, clearly make the decision, but go in incrementally. And please don't underestimate the value of your workforce and use this moment to actually evolve your working environment so people can actually apply the things that they learn as well, which is also a big sailing for leaders today. But yeah, I think there's like some very high principles that leaders need to think about with all of this. And, you know, it's not, it's not easy for them because, you know, you and I appreciate this, that AI is just one big smelly pillar that they have to deal with on a daily basis, along with the regulator and many other things that are driving their balance sheet. But this is a big one that would either make them successful or erode it over time.
- [72:12]
B
I really appreciate that insight. Thank you, Michael, for joining the show here today. For all the insight that you've shared, and I really appreciate it. I think there's a lot there for leaders to learn from and to consume and help them make better decisions and ask better questions.
- [72:29]
A
It's a pleasure.
- [72:32]
B
Most viewers don't know this, but Digital Disruption is developed by Infotech Research Group, a leading advisor to technology leaders around the world. If that's not you, you don't need to care. So skip ahead and enjoy our content. But if you are a technology leader, Infotech helps IT teams get projects done faster, better, and at a lower cost. Infotech provides unlimited access to practical tools and expert guidance you need to execute at a fraction of the cost of traditional consulting, no matter the project. From AI strategy to cybersecurity to vendor negotiation, Infotech has you covered. Check it out at the link below. And don't forget to like and subscribe.