Whose interests are you serving?
On digital sovereignty — Part III
At the end of Part II, I left you with the question that“you must decide whose interests you’re serving”. At first glance, this seems like an easy question to answer i.e. “ours!”. Alas, if only life was that simple.
To begin our journey into Part III, we will start by comparing nations, looking into the concept of culture, asking some question on culture, mixing in a bit of AI and policy before summarising. So, grab hold of your hats, as this promises to be a rollercoaster … weeeeeeee!!
Comparing nations
In 2014, I was deep into mapping economic and technological spaces. Those questions on digital sovereignty had started long before my exercise at the DVLA. One of things I had undertaken was not just the mapping of industries (often used in M&A discussions or investments, particularly shorting) but the comparison of nation states.
These were mad, year long exercises involving mapping out many industries and the wider economic space. Of course, parts were easier i.e. mapping out a discrete landscape such as drones or trains was relatively trivial compared to examining government, financial and manufacturing capabilities. On top of this, I would quickly find myself in discussione on fairly nebulous terms such as values, optimism, mobility and existing beliefs.
It was after one of these efforts (examining China and the US) that I found myself in a conference room in Washington DC presenting to some powerful figures. Alas, what I presented did not go down well. I explained how across a range of industries (a few example are given in figure 1) that the US would lose or already had lost its dominance to China. The message did not land well.
I had tried to show maps, show how China was playing the game, the use of a mixed economic model with directed investment and how this had been successfully developed over decades by a continuous program (that combines everything from SEZ to education to high degrees of situational awareness) as could be seen by simply looking at import/export ratios across supply chains. The latest data I had at that time was 2013 (see figure 2) and clearly China was not only producing more high end goods, it was shipping more to the US rather than the US dominating those markets.
I tried to explain mechanisms of how to turn this around, how to use more directed investment, the important changes that were needed within education systems and investment in understanding supply chains. I warned about tariffs and trade wars and the more I tried to explain, the more evidence I presented, then the bigger the hole that I was digging became.
You should remember that this was over a decade ago.
The message landed so poorly in that gathering that I was quite bluntly told that I was talking nonsense, that “Silicon Valley would out innovate China” and I was ushered out of the room. As my colleague at the time explained, that did not go down well.
On the upside, I kept pondering why did they believe “Silicon Valley would out innovate China”? Sure, there was racism in there but there was also this belief but where did it come from? I tried to ignore this whilst I started examining Iran versus India, but it was during this process I kept head butting concepts of beliefs and optimism. Eventually, that sent me spiralling off into a journey to explore the concept of culture.
Getting a handle on culture
I would constantly be told that Silicon Valley would keep the US ahead of the game because it had a more innovative culture and that “China just copied”. This made no sense to me. I had visited and studied China back in 2014 to 2015 and what I saw was a highly optimistic and innovative environment, rapidly developing. It was breathtaking. The “percieved” wisdom made no sense at all. But then the problem was explained to me — “I didn’t understand culture” and “it’s all about having the right culture”. Those lines sent me off the edge into a rage.
After I had calmed down, I decided to explore the subject because it was a fair cop — I didn’t have a handle on culture. I thought I’d better start with a definition of what culture is. So, I bought every book I could find and started to read. What I found surprised me.
In 1952, the American anthropologists, Kroeber and Kluckhohn, critically reviewed concepts and definitions of culture in the academic world and compiled a list of 164 different definitions. There is no consensus amongst anthropologists on what culture is. At best, we can say that culture contains many things from experience, learned behaviour, knowledge, meanings, relationships, hierarchies (i.e. power structures), capabilities (i.e. skills), values, possessions, aspects of belief (i.e. ethical position relative to others), principles and attitudes. It has many layers from the individual to the group to the organisation to the nation. It has many perspectives which may vary with context and many attributes appear emergent, as you would find with many evolving processes.
However, this lack of clarity and wealth of complexity doesn’t stop a business press littered with publications promising readers how to build “the right culture” as blessed by some guru and assigning almost mystical properties and powers to it with memes such as “culture eats strategy for breakfast”. It felt like I had walked into a world of cults and all I wanted was a simple definition. Fortunately, I was saved by the work of a truly wonderful anthropologist — Margaret Mead and her single line that “Language is a discipline of cultural behaviour.”
Margaret had noted that language was part of the cultural system. Now, that’s fascinating because you can never create a complete and true understanding of a system within itself, you have to step outside. If language was part of a cultural system, then I was going to have to find another way to look at culture … and, of course, I had maps.
But how? It’s all around us but we can’t describe it and I knew I would somehow have to gather together diverse viewpoints to try and elucidate what it is. Well, I often used maps in conflict resolution because they are great tools for getting conflicting sides to discuss a problem without confronting each other. By pure chance, the Brexit debate was raging in the UK. So, I was able to get two groups of people — Brexiteers and Remainers — who literally couldn’t talk to each other in a civil manner and got them, through the use of a map, to find a common understanding of what culture is. This is shown in figure 3.
Ok, it’s worth a bit of explaination. The axis at the bottom uses concept, emerging, convergent and accepted which has the same meaning as genesis, custom built, product and commodity. They are just different labels for different classes of capital. A number of the components are pipelines (square boxes) which means they represent many components, all evolving i.e. we don’t have one value, we have many.
Starting from the top, when we are talking about cultural systems I had to start with this idea of We (as in our group) and Me (as in an individual). Though this starts off from different places, they rapidly converge into ideas such as power, collective, values, behaviours, safety and a sense of belonging. There are lots of loops in this map — those loops can be stabilising or destabilising.
Even simple things like behaviour is a result of interaction between doctrine (our principles) and values (our beliefs) i.e. we might share a common principle of “serving our users” but if one group’s belief is that “people with green eyes are more important” and another group’s belief is “people with brown eyes are more important” then their behaviours are going to be different, even if they share the same principle.
I won’t go into mapping cultural systems (or the messy space of politics) other than to note that just as with territorial, economic and technological landscapes then you can map it out (see Part I — Sovereignty and Landscape)
One thing that did however fascinate me on the map was how our collective memory (which we often describe as art) i.e. the heroes, rituals and symbols of our culture can influence our values. That art can represent itself in painting, in statues, in stories, in music and in video games (yes, they too are a form of art). All of these can help shape our culture.
Take the UK story of Robin Hood, which if you don’t know is all about the heroic efforts of a lawman (the Sheriff of Nottingham) in trying to prevent ruthless gangs of outlaws (led by an arch criminal, a hooded man called Robin) who are running amok and terrorising law abiding citizens … oh wait, that’s not how we tell it. Instead, we tell a story of social responsibility — of giving to the poor. How the story (the art) is told matters. Change this, and you change our values.
One thing I will note is that this is a description of culture but you’re not part of any one culture, you are almost certainly part of many — your family, your local sports club, your work, your nation — and all these cultural systems interact.
Art as social guardrails.
So, why Art? Throughout history, art has been used to change the values of societies — inspiring peaceful change, to bloody revolutions to monstrous tyrannies. You don’t think organisations like Hezbollah produce AAA video games because they’re “gamers” do you? Most collectives conduct a form of social warfare through art, some are simply more sophisticated in their approaches.
And art matters because it encapsulates and shape our values. In figure 4, I’ve provided a fairly simplistic analysis of important stories within different national collectives and the values they are associated with. Of course, you have to be careful with stereotypes here, and have to try and understand the environment those stories exist within but it gives a sense that we often have common values underpinned by different stories.
That doesn’t mean every culture is the same, the US and UK are far more individualistic whereas both China and UK have strong signals of social responsibility and all three share strong values of equality and justice.
Art (i.e. the collective memory of heroes, of symbols, of rituals) is not only a path to changing a cultural system (i.e. alter the meaning of the story and you change the value), it also creates inertia through the social capital we have invested in this. The ideas of redistribution and Robin Hood are difficult to escape in the UK. If you retold it with the Sheriff as the hero then people would look at you oddly.
Unfortunately, this is where the real danger of AI and the new theocracies that I discussed back in 2023/2024 come into play. Justine Tunney wrote a lovely tweet on “Google Gemma 27b is a techno-libertarian” (see below) — well, what if it is? How would you know? We already know these AI systems show incredible market bias over societal. Ask yourself, how much damage could be done to a society if you forced school children to learn with these systems when you didn’t know what values / beliefs they had been trained with and you’re not teaching those same children to critically think?
Before you ask, I mapped out the education system from multiple perspectives with teachers, professors of education etc in 2022 / 2023. We used the maps to look at where we need to invest (both from a societal and a market benefit). Those are not aligned (see figure 6).
Don’t blame the teachers though because this is despite the best efforts of teachers. Alas, the purpose of education is mostly about producing useful economic units (supporting Government needs for a market) and meeting the needs of the educational establishment. We came to this conclusion because the maps only made sense when we mapped from those anchors (Government needs, Educational Establishment needs) rather than societal or student needs. It was tough pill to swallow for people working in this field. In the world of AI we should be investing in critical thinking, it should be a formal subject taught to every child. It isn’t though.
AI and policy
At this point, I’m going to come up with a crazy idea. The greatest victory is not in conquering your competitor through conflict, but in making them become you. In this world of competition (of conflict, of co-operation, of collaboration) with others, what if we could bypass that cultural inertia in other collectives? What if we could get out very own Wormtongue into the heart of other nations and other organisations? What if we could persuade those others to listen enthusiastically or even better, not to be aware that they they are listening to our stories at all? What if we could just make others accept our values, to make their decisions based upon our values and become more like us? That’s very Sauron. Of course, they’d never go for it … would they?
There are four images I want you to look at, the first is from that DVLA exercise in Part I where we discussed the importance of having borders around training data (i.e. the simulation models for the AI).
What we were saying back in 2015, was that for reasons of national sovereignty we needed to protect those spaces, we needed a border around it because we thought it was important to protect our sovereign values. However, to be fair, though we could articulate it on a map, we couldn’t really explain why it was important other than through the narrative of the participants. I had an early form of cultural mapping but hadn’t gone through any major experiments, such as the Brexit vs Remainer groups. So, as usual, I was a bit reluctant to dig too deeply.
Of course, once we had a cultural map (see figure 8), we could talk about how those values underpin behaviour and stability in our society. Top of the map (i.e. the anchors) are expressed in terms of “We” and “Me”. This is not an either or, as all cultures are a bit of both except for those really unstable ones (i.e. communism and libertarian laissez faire ideals).
Those concepts of “We” and “Me” are often expressed in terms of inclusion and exclusion (as we discussed in Part II — Societal versus Market Benefit). We need both, exclusion for a functioning market (i.e. trade) and inclusion for opportunity (see figure 9).
Unfortunately, as we also noted in Part II, the AI systems available tend to have a bias towards investment in the market rather than society (see figure 10) i.e. they are a reflection of market data and I suspect the US culture they were trained in.
AI is that perfect Wormtongue and in many cases it’s your very own “Ayn Rand” whispering sweet tales of objectivism and sovereign individuals. This is why it’s really important to understand what it is trained on and why training data must be open. This is much more insiduous and dangerous than Frontier AI and relived youthful nightmares over Skynet. And people know it … just say “I think we should force the use of DeepSeek everywhere because it’s more open” and watch all the wannabe techno-feudal lords start moaning about China’s influence and dangers.
Alas, we don’t tend to look at these landscapes when we talk about digital sovereignty. We instead have discussions about owning stacks or data or … well, frankly, they are as poor as the discussions on culture. We should be looking at these landscapes. If we did, we should be concerned about using any of these AI systems in the creation of policy at a Government level unless we’re ok with potentially importing a more market focus and different cultural values into our system. If you look at Silicon Valley or the US in general and think “We should be more like that” then fine. If you exist in a different cultural system then you might want to pull the brakes on this a bit until you truly understand what these systems are trained on.
Summary
In Part I, we discussed the importance of mapping the landscape for soveriengty — not just the territorial but the economic, technological, social and political. In Part II we discussed what you do once you have the map i.e. you invest. But you must decide whose interests you’re serving. In Part III, I hope I’ve shown you that such questions lead us into deep cultural landscape of competing values, behaviours and collectives. These are the sorts of discussions that should be happening around digital sovereignty.
Final Notes
There’s a fair point to be made, before we wrap up, about tone and limitation. While this essay series has aimed to be informative, it hasn’t pulled punches and in doing so, may at times have trampled nuance. The conviction with which I argue for mapping, and for asking hard questions of our strategic intent, may come across as dogmatic to some. That’s a fair criticism. But I’d rather be accused of being overly direct than contribute another bland report that nods at complexity and quietly steps away from consequence. If that makes some uncomfortable, good. Comfort rarely produces clarity.
That said, the intention here has never been to dismiss market approaches, technology, or even traditional strategy, but rather to point out where they fall short and more importantly, what they overlook. The series is not a manifesto against the market; it is a call for balance, for remembering that not all value is commercial, and not all progress is profitable. The market does what it does well: it allocates resources, it finds efficiency. But it does not think about justice, inclusion, or cultural survival nor should we expect it to. That’s our job. Strategy, if it’s to be worthy of the name, must do more than follow the money. It must also follow the map, and decide — clearly and deliberately — whose future we are building.
Finally, my rant is … over.
….
On digital sovereignty series.
Part I — Sovereignty and Landscape
Part II — Societal versus Market benefit
Part III — Whose interests are you serving?
