Another day, another exhortation to join an “ecosystem” that’s anything but. I could pick a hundred examples, but one that recently caught my eye was an ad placed in the Financial Times by the Singapore stock exchange, SGX Group, promising “multiple growth avenues, one trusted ecosystem”. SGX wants companies to list on its exchange rather than, say, the Hong Kong one which has more or less the same exclusive offer. SGX promises “access (to) Asia through our trusted ecosystem anchored in Singapore.” Ecosystems can be a lot of things. ‘Trusted’, by which they mean centrally policed to achieve defined, lower-risk outcomes, is not one of them. Calling built environments ‘ecosystems’ is common everywhere from financial services to supply chains to – quite a stretch, here – a retirement living complex. But it’s most often used in the tech world to describe the relations between software, services and hardware typically owned by a single company. For example, it’s how Google describes everything that hangs off the Android operating system. These kinds of proprietary and deterministic architectures are called ecosystems so often that we’ve stopped noticing. And that’s kind of the point. We need to start seeing this metaphor again, and what it’s hiding in plain sight. But first, a reminder of what an ecosystem actually is.
An ecosystem is a set of unbidden organisms and the physical environment with and in which they interact. It’s constantly evolving, and the real interest, value and drive for change all come from the emergent properties of the relations between its many parts. An ecosystem is not the plaything of a pampered princeling, like Meta, but a set of living, striving things, both competitive and cooperative, and the place they live. The two kinds of system are almost impossibly different. One is biological, the other technological. One is complex and adaptive, the other only pretends to be.
Why are the maddening, built environments that certain investors and their pet CEOs want us to spend our lives inside called ecosystems, when they’re the very opposite of anything truly alive? It may be that to people with a radically simplifying and totalising mindset, a tech maze governed by decision tree really does seem wild. Adding a few menu options makes a walled garden into the savannah, to someone who believes freedom is a luxury product. But whatever it says about the denuded imaginations of system-owners, the use of ecosystem as a pure marketing term has a strong halo effect. ‘Ecosystem’ implies competition, dynamism and choice, even as the systems it describes are restrictive and monopolising. It invokes our natural world, even as it conjures a stampeding flight from it. As a PR strategy, the misused metaphor is a sleight of hand. But it’s more than that.
By subtly invoking both competition and containment, ‘tech ecosystem’ enacts an inner conflict that many in those companies may be only unconsciously aware of. They’re driven to be global, but in a rigidly parochial, imperialist way that serves the cultural values and geopolitical interests of just one country. They need constant acceleration, but only along rigidly determined axes. They invoke a sense of limitlessness while herding people – ‘users’ – into feeding lots. They want to be the whole world, the ultimate everything app, but also to edit out unprofitable activities and externalise their own negative consequences. They say ecosystem. They mean snow-globe. It’s not at all clear if they understand these are different kinds of thing.
Calling your company’s assets an ecosystem in this slyly limited way is a trick to divert attention and responsibility. It redefines what’s inside the ecosystem – profit, control – and what’s outside – the costs to all of us. It’s also a Trojan horse for the insidious idea that technology systems are sovereign and discrete, whole worlds unto themselves, less subject to the ties and constraints that bind us all. Yes, I can see why the owners of these constructs wish us to think of them as ecosystems.
There are two responses to the ubiquity and insidious violence of this metaphor. The short one: tech platforms and proprietary software environments are not ecosystems, so don’t call them that. Call them built environments, i.e. designed, rules-based systems that explicitly structure interests to secure specific intended outcomes. It does no good – for journalists in particular – to transmit the suggestion that a walled garden is the same as a living forest. That an app market-place is the same kind of thing as an open protocol. We don’t just serve the interests of system-owners when we repeat the pretty lie. We shut down an essential way to imagine alternatives. So what if, every time we read ‘ecosystem’, we instead say ‘plantation’?
A plantation is a hierarchical, exploitative monoculture. In agriculture, it’s “an .. estate, generally centred on a plantation house, meant for farming that specializes in cash crops, usually mainly planted with a single crop”. Google’s interlinked extractive systems are plantations whose single crop is data for ads. They’re designed environments; their parent company, Alphabet, a conurbation of control. Amazon’s warehouses and proprietary, spyware-based web services arm are plantations, with all the connotations of colonialism and coerced labour the metaphor carries. Its marketplace for sellers binds them into literal share-cropping, an exploitative type of tenancy that binds farmers to land they’ll never own. Facebook’s hard-right political pollution of each country it operates in echoes the ecological damage of how single crop, intensive plantations do great secondary damage, destroying habitats and biodiversity, while generating vast profits for a tiny ownership class. We already have metaphors that describe exactly what these companies do. We need to use them.
Accepting the ecosystem metaphor for technology firms distorts our discourse. It implies there is dynamic competition and choice when in fact we live in a world of infrastructure chokepoints and application-level global duopoly. The way we currently ‘internet’ is based on a single mode; surveillance and advertising. It’s not just that we have little choice, as individuals, but that the organisations ostensibly battling for our custom (or servitude) compete only on marginal operational differences in the same basic strategy. They have the same business model, the same financial backers, and the same drive for what they call growth and scale, but which is merely old-fashioned monopoly. The Internet as most people experience it is not even a walled garden, but just another kind of industrialised agriculture. Calling it an ecosystem subliminally green-washes the damage these and other less well-known companies do, including to our physical environment.
Metaphors shine a light (see what I did there?) on an aspect of something that might otherwise remain unseen. But they can also flatten or even subtract meaning, or just outright lie. Like with accusations made by toxic narcissists, the ecosystem metaphor unwittingly reveals that the systems it describes do the precise opposite of what it implies. It’s a monumental tell. I’m struck by how, at the precise moment of this civilisation’s greatest harm (yet) to the actual, real-life ecosystems we depend on, the term has been appropriated to describe not just merely its own opposite, but the purest form of their destruction.
The second response to how the term ecosystem is used by its enemies to describe its opposite is to see what else the metaphor might reveal. If it didn’t work so well, both for those who employ it and we who receive it, it wouldn’t be used so much. I think it also expresses something many of us deeply desire but can barely articulate. It is seen everywhere but experienced almost nowhere. It’s surfaced so often to suggest but instantly cauterise our longing for another way to build and experience the internet. But that’s useful. We can use it both to see what’s really there and to suggest what might yet be made.
The people who originally came up with the term ‘digital ecosystem’ knew this, and defined it as “a distributed, adaptive, open socio-technical system with properties of self-organisation, scalability and sustainability inspired from natural ecosystems.” That is clearly a very different kind of environment to the dominant platforms! There are three key ways in which our current tech “ecosystems” differ from real ones. I describe them here not to keep making the trivially obvious point that what we’re currently offered are not actually ecosystems, but to suggest what a real digital ecosystem would be like. We are so far down the road of ‘there is no alternative’ kinds of systems –political, economic and technological – that it can seem near impossible to imagine alternatives. But they really exist. They’re not just metaphors. This is just a beginning.
Diversity
Functioning ecosystems tend towards biological abundance, both in quantity and diversity. In contrast, plantations have high volume but concentrate on only one or two things, often with secondary damage to the actual ecosystems nearby. However, since the 1970s, the decline of biological abundance has accelerated. Insect biomass in Germany down to a quarter of what it was then, US farm birds down by 3 billion, global freshwater creatures down to 15%. The Internet’s diversity has contracted, too. Designed as an open and emergent communications architecture, the Internet in the US now has two, maybe three wildly dominant content distribution networks, and just a couple of DNS providers. (The emergence of these network-level middle-boxes at all, and how they concentrate and ostensibly speed up traffic is itself an admission of failure, and their evolution into oligopolies an indictment.) From its invisible infrastructure to the consumer-facing internet that we can all see, our network and information systems are becoming monocultures.
But the antithesis of biological abundance isn’t just biological paucity. It’s the excess growth of one thing at the expense of all others. Looking at the internet like a real ecosystem reveals that the torrents of disinformation and logarithmic rage cultivated by social media platforms uncannily mimic the lurid, obscene algal blooms in rivers and estuaries force-fed by agricultural effluent. Tech platforms push the growth accelerant of ‘engagement’ just as intensive farming sprays fertilisers packed with far more phosphate and nitrogen than soil can absorb, and with little care for the run-off and its consequences elsewhere. Social media literally maddens people. We know that. Thinking of it as a cultivated system that pollution has driven into eutrophication – when algal blooms consume all the oxygen and kill off other species – reveals more of what’s really happening. It lets us see the phenomenon a little differently, and also reminds us that this kind of problem isn’t new. Nor are its solutions.
At its core, environmental regulation pushes costs that have been externalised right back at the entity that’s produced them. We know how to deal with effluent. We know how to stop acid rain. Building the political coalitions to punish and disincentivise polluters is still incredibly hard, but people manage it all the time. Tech activists have much to learn from our environmental brethren, and perhaps some things to share, too. But first, we need to recognise that our information sphere is a eutrophication dead zone. It’s a literal disaster. That framing communicates the urgency of the problem and the scale of its reach, but also puts it in the category of problems we already know how to solve.
Value
In a real ecosystem, value is generated by the relationships between different parts of the system. The relationships between plants, animals, fungus, etc. and their physical environment have so many more facets than we can easily capture, each with its own inherent value. But in a tech plantation, value is something to be extracted from the simplified, observable and manipulable parts of the system by the people who design and own it, typically tech bros and their VC backers. They manipulate how the parts of – or people in – the system interact so they can maximise the data generated, ads served and revenue extracted. That much is obvious. But looking afresh through the lens of a real ecosystem refocuses our attention on where and what value is, and to whom.
In a platform system, value is generated by forcing the constituent parts to interact with each other in a radically simplified way, and funnelling the revenue generated to a tiny number of system owners. (The costs – from the bruised lives of victims of image-based abuse, to mortally wounded democracies, to inciting and equipping actual genocide – are externalised.) But the value of an ecosystem is generated in one to one and many to many interactions, thrown off willy nilly to the organisms and environment of the ecosystem itself. The ‘value’ generated moment to moment in an ecosystem is multi-valent and often far from obvious. To use the Yellowstone wolf rewilding story many of us know, reintroducing a keystone or functional species caused positive, unpredicted and cascading changes in the relations between other organisms. Wolves preying on elk meant the elk moved around more and ate less river-edge willow, which meant more willow for beavers to build dams, which meant more stable river flow, more and healthier fish, and bigger habitats for water-adjacent songbirds. The profound, indirect and cascading impacts of wolf reintroduction surprised pretty much everyone. The ‘value’ in a real, functioning ecosystem is not a simplified, hierarchical, bro-tastic chain of predation, but a set of multi-faceted relationships where the smallest, plainest organisms can be just as important. Or, as Robin Berjon has put it, “in an ecosystem everything is infrastructure for everything else.” The relations between the different species are where the ‘value’ is. Food is not a radically legible chain. It’s a web, with drops of glittering dew on each thread refracting sunlight in every colour, every direction.
I love how the Yellowstone wolf story works in two directions. It highlights the importance of ‘capstone’ or functional species like a large predator, but shows this animal’s significance is how it sets off a multiplicative ripple of changes in the relations beween scores of other kinds of organism to subtly to change the ecosystem as a whole. It starts with an almost hyper-masculine fixation on the ‘apex’ predator and its kills, but – and we’re still only in the middle of that particular story – changes quickly to being about grazing and walking, swimming and singing, fear and possibility. It works as a fable both of agency and chaos, of taking directive action while holding space for myriad unanticipated trickledowns. It’s an anthem to the possibilities for acting, and the wholly unpredictable delights in store when we do.
What would an ecological value generation model look like on the Internet? Well, many would say it’s what we had in the early, decentralised days of the Internet and well into the world wide web, before the current business model monoculture took hold. But the multivalent value generation of the food web is still with us. It’s just, as the man said, unevenly distributed. You can see it in the traffic co-opetition model of many Internet exchange points (IXPs), the federated instances of Mastodon and the like, in community networks and municipal ISPs and indigenous connectivity networks, in standards work to decentralise internet functions. It’s even in attribution model of the photo I used at the top of this piece. And it’s increasingly in people’s consciousness. The hostile takeover of Twitter has prompted many, not just tech policy specialists, to now understand and demand social media interoperability and the true autonomy and choice it might bring. People now get this, in mind and heart, in a way they did not just a month ago.
One way ecologists conceive of ecosystems is as flows of food energy and material. An ecosystem in more or less stable equilibrium will have a steady amount of biomass (the total mass of living matter), with the same amount of energy circulating from one moment to the next. Roughly speaking, it’s a cycle; a more complex version of the yin and yang of how the equations for photosynthesis and respiration fit together. From this perspective, our tech platforms don’t look like an energy cycle. They generate a huge amount of energy – i.e. data and money – but systematically suck it out. Data is hoarded; almost none is shared with the people it’s about, and the only transfers are when it’s sold or rented to other wealthy actors, or shared with governments. Money is channelled out of local economies and tax systems to be parked in nation-sized tax dodges or jet-sprayed at tiny, dysfunctional locales such as San Francisco. Seen as energy systems, tech platforms are simply extractors who shove a violently distorting amount of catalyst into what were once organically occurring reactions, and capture most of the energy created to spend – or hoard – elsewhere. It’s not an equation. It’s sure as hell not the circle of life. It’s simply a subtraction.
Outside the maddening, extractive current system of how we currently produce and reproduce the Internet, people are not a chain gang, not a one-way, extractable resource. We’re not a commodity. We’re multi-faceted. We give and gain and often unwittingly generate all sorts of value – monetary and otherwise – in our interactions with everyone and everything we touch. We cannot be reduced to what someone can extract from us. And none of us is intrinsically worth more than the others. We feel and act and interact like the elements of an ecosystem, because that’s what we are. It stands to reason that shoving us into a plantation brings out the worst of us. It also implicitly puts most of us into a subordinate position, based on how much value we produce for the system owners. It takes looking at the Internet as a real ecosystem to see the weirdness of how its energy is generated and where it’s flowing, and how distorted it is that much of the value it generates is funnelled out by owners, not bounced around between the different elements of the system, i.e. us.
So much of human life and worth is missing from the stripped down and hyped up way we currently internet. There are ways we can fix that, and first among them is halting the drive to re-engineer our societies and economies to fit into these built and extractive systems. But there’s also a ‘not seeing the wood for the trees’ side of it. The way we currently internet is made maddening by the perceived need to pay for it with ads. At the heart of our current system is an absolute monoculture of business models; we trade advertising and the data it depends on to provide us with content and subsidise our connectivity. That’s it. That’s literally the sole business model for the internet. Billions of people on the planet, trillions of unique ideas, but only one acceptably “scalable” concept to connect them. The vast system this single business model built extracts far more money than most of us can imagine (and will continue to do so even now the internet economy is contracting, just at a lower rate). We might get the same connectivity, content and services – if not more – by keeping the value generated by the internet system in the system. We’ve just never yet had the chance to try.
Control
In the most trivially obvious way, tech platforms are designed systems of hierarchical control, while ecosystems are sets of complex, emergent, multi-valent interaction where each organism is held in check by the others and their overall environment. Ecosystems don’t police themselves, nor do they have intentional checks and balances. But with all that food and energy flowing about, there will always be a new, competitive niche just begging to be filled. The first computer simulation I ever saw was a simple, dynamic graph that ran on a ZX Spectrum 48k, called ‘rabbits and foxes’. (It was designed to show how those tiny but versatile machines could run differential equations to model predator-prey interaction.) The x-axis showed population numbers of each species, while time ran along y-axis. It was quite a slow simulation, if my memory is correct. The rabbits bred like … rabbits, so their line tracked up, point by point, followed by the foxes who ate them and thus also increased. But then the many foxes would eat too many rabbits, so the rabbits tracked down again, followed – as swiftly as a 1983 era personal computer could manage – by the foxes. This undulating duet travelled on indefinitely as a soft-arced double sine wave. It left a deep impression on me. Firstly and most obviously, that the predator only looks like they’re in charge. Also, that nothing lasts forever. And perhaps, though this came much later, that at a big enough scale predation starts to look more like competition, starts to look more like cooperation.
In contrast to a functioning ecosystem, app stores are centrally controlled by a single power. Innovation runs along pre-set rails. Apps compete with each other, but only on criteria set by the system owners. Users choose only from what the controllers make available. The money, data and power generated in the app store flow disproportionately back to the owner. It’s emphatically not an ecosystem but a hierarchy that permits ritualistic competition to drain off the energies of potential competitors, like tame opposition parties in places like Singapore, or the jangled extremism of parliamentary factions in Russia. Functioning ecosystems have three types of fundamental interactions between organisms; competition, predation, and what we might call cooperation, but in the absence of each organism having a theory of mind about the others it interacts with is more accurately termed facilitation. It’s clear that competition in technology plantations is limited and performative, and cooperation is similarly constrained. Just ask a tech worker trying to unionise or make real interoperability work. But these systems don’t just use fake competition to hide the absence of the real thing. They obscure how the primary way the system owners interact with us is predation.
We don’t live in competitive tech systems but in primarily predatory ones. And unlike in an actual ecosystem, where predators bounce up against limits – be they competitors, territorial conditions or the adaptive behaviour of prey – our platform predators have few limits. They’re richer than most governments, so rarely face competition from other types of predator. They use their scale, wealth and power to shape the environment in which they operate – through lobbying, withholding tax to denude public services of resources, and then privatising them, and by eliminating alternatives or opposition through acquisitions, the destruction of media, the corruption of the academy and the purchase of tame non-profits or their erstwhile employees. Perhaps more importantly, they shape the mental world of everyone who might some day oppose them or build an alternative. No apex predator could dream of subtly convincing her prey that it’s better for them to simply not run away. I guess that, fundamentally, is what makes us human.
It’s not the wildly disproportionate amount of power platform that controllers hold that’s most dangerous, it’s how they wield it to prevent us imagining alternatives. That’s why the ecosystem metaphor must be shot down everywhere it pops up. We don’t live in a technology ecosystem. It’s more like a zoo crossed with a detention centre. Or, to use an image that also conveys the beautiful lies that hold us in place, we live in a snow globe. It’s a pretty albeit ersatz version of what a town is, and its limits aren’t visible until you run right into them and smash your face on the glass. Who’s holding the snow globe, peering in?
Recently the FT ran a chilling piece of business reporting about Apple in China; ‘Apple trades acquiescence for access to factories and consumers in China’. Apple is the ultimate closed system in tech. (I should know. I’m writing this on one.) Apple’s manufacturing is concentrated in a single region of China, and the company does what it has to do to keep that working. If that means hosting all customer data where the government can directly access it, sure. Shutting down airdrop between anonymous machines when people started using it to organise protests? Fine. Removing apps, pre-emptively censoring people’s words, banning encrypted communications – absolutely. A predator instinctively corners its prey, shuts down any chance of escape. The only thing a predator fears and respects is another predator. Apple’s closed and controlling environment fits hand in glove with its global enabler, the CCP. The article concludes with the words of Nathan Freitas, a developer of secure and decentralised communications systems:
“Apple’s vision of a controlled, locked down ecosystem … maps into the same vision, the same control that the Communist party wants to have in China. They see eye to eye on what, for a harmonious society you need. It’s just one is a phone ecosystem, the other is a nation.”
Conclusions
What might it be like to live in an actual digital ecosystem, one that shared the key characteristics of functioning biological ecosystems? Functioning ecosystems are defined by competition, predation and cooperation (or facilitation). Each of these is a building block to complex and adaptive communities. So, it wouldn’t always be pretty. It would be messy, and sometimes we would lose. Sometimes, but not all the time. That’s the difference.
Taking up the challenge to think of technology systems and their version of the internet as an ecosystem reveals that ecosystem to be profoundly damaged. It also focuses our attention on the arbitrary lines these systems have drawn between what’s inside them and what’s outside. Costs and harms are externalised, but value flows out, too. For an ecosystem, this looks aberrant and downright weird. But as a civilisation we’ve already come up with ways to tackle these kinds of problem and the people and organisations that make them. We know how to re-internalise costs. We know how to break monopolies. We know how to punish facilitators of harm.
Ecologists also are beginning to know how to create the conditions for real-life ecosystems to repair themselves. Sometimes that means reintroducing functional species, especially large grazers. Sometimes it means ripping out invasive species – in much of Ireland, that means rhododendron, a beautiful flowering shrub that permits nothing to grow in its shade – and letting shrub and then native trees re-establish themselves, which they do surprisingly quickly. (I’m always wary of mentioning ‘invasive species’. It’s a biological analogy poised to sprint out of context and into a deeply unsavoury metaphor.) The first step is to see things, even pretty, useful things, for what they really are. George Monbiot points out that the beauty we perceive in landscapes denuded of trees – for example, Dartmoor or the Scottish highlands – is a falsehood unless we read the clear, undulating lines of naked hills as landscapes whose saplings are repeatedly ripped out. Ecological repair requires us to put away myths and metaphors, and see bare biological facts. Both ecological and technological repair need us to recognise the ‘shifting baselines’ of impoverishment we think of as normal may just be a generational adjustment to a worsening situation we need to urgently arrest.
For technologists, the challenge is to resist the taught aesthetic and even obliquely moral preference for centralisation, simplification and control. These values serve the interests of owners obsessed with permanent, eutrophic growth, but have no fundamental, over-riding value of their own. Computing and engineering have a monoculture of values that needs to be cut down so that others can grow. They have so much to learn from other fields. The way in is through the childlike curiosity and delight that brought us all into this space.
Metaphors work best when we consciously use them as tools to look with, rather than as the whole thing we’re seeing. Tech systems – plantations – are of course not ecosystems, and we can refuse the invitation to uncritically see them that way. But we can do more; we can think ecologically about how to fix the harms of these controlling systems. And, vitally, we can start to fulfil the broken promise of what we truly need them to be.