The Information Bloom
Abstract
The information revolution — and AI is its latest phase — is like a nutrient flood in an ecosystem. Algae are the organisms best placed to exploit nitrogen, and when nitrogen floods in they bloom rapidly, consuming the oxygen and crowding out diversity. In 1971 Herbert Simon warned that a wealth of information creates a poverty of attention. Today that poverty is acute. But the very quality that makes us bandwidth-limited also makes us irreplaceable: we carry the knowledge that comes from being mortal and vulnerable. This talk explores the embodiment factor — the million-to-one gap between how fast we communicate and how fast machines do — and argues that human culture, like ecology, carries complexity that cannot be scaled or uploaded. The challenge before us is to ensure the information bloom does not destroy the diverse ecosystem of human voice and face.
There is a phenomenon ecologists call an algal bloom. Algae are the organisms in our waterways that are best placed to exploit nitrogen. When nitrogen floods into an ecosystem, from agricultural runoff, from fertiliser-laden rainfall, algae do what they are optimised to do: they grow. Rapidly, voraciously, they spread across the surface of the water. To the casual observer it can look impressive, even beautiful. But the bloom is catastrophic. It consumes the dissolved oxygen. It kills the fish. It crowds out every other form of life. The very success of the algae, their perfect adaptation to exploit a suddenly abundant resource, destroys the ecosystem that once sustained that diversity.
I want to suggest to you that we are experiencing something very similar in the domain of information. And that the question this conference poses: how do we preserve human voices and faces? Is precisely the right question to be asking, and asking with urgency.
In 1971, the American economist and computer scientist Herbert Simon wrote something that should have alarmed us long before it did. He observed: “What information consumes is rather obvious: it consumes the attention of its recipients. Hence a wealth of information creates a poverty of attention.”
A wealth of information creates a poverty of attention. Simon saw it coming fifty years ago. The information revolution, and AI is simply its latest and most powerful phase, does not simply give us more. Like nitrogen flooding a waterway, it can overwhelm the ecosystem of human attention, human culture, and human voice. The organisms best placed to exploit the flood will bloom. What we must ask is what they crowd out.
Locked In: The Limits of Human Bandwidth
To understand why human attention is irreplaceable, I want to introduce you to Jean-Dominique Bauby. Bauby was the editor-in-chief of French Elle magazine. Bauby was a man whose life was words, ideas, communication. That was until he suffered a brainstem stroke in 1995 that left him entirely paralysed, save for one eye. He became what medicine calls “locked in.” His mind was perfectly intact. His thoughts were free, vivid, imaginative. But his ability to communicate was reduced to blinking.
Bauby wrote a book after he became locked in. He blinked at a board on which a nurse pointed to letters in order of their frequency in French. Each word took two minutes. The whole book took ten months of four hours a day.
One word every two minutes. Using the information theory developed by the mathematician Claude Shannon, we can quantify this precisely. Shannon showed how to measure information in bits. One bit is the information contained in a fair coin toss. He estimated that an average English word carries about 12 bits. One word every two minutes means Bauby was sharing roughly 6 bits per minute.
At a typical speaking rate, a person communicates around 2,000 bits per minute, roughly 300 times faster than Bauby. That feels like an enormous difference.
But here is the truth that should give us pause: relative to a modern computer, you and I are in the same situation as Bauby.
A computer communicates at around 60 billion bits per minute over a standard network connection. Compared to a machine, we are all locked in. We are all blinking at the board, one letter at a time.
This has been true for decades. What is new is that we have placed machines operating at those speeds into the very centre of our information environment. Those machines are between us and our news, between us and our data, between us and each other. The machine has inserted itself into the ecology.
Bandwidth as the Source of Humanity
But here is the crucial point. The fact that we are bandwidth-limited, that our attention is constrained, that we are “locked in” relative to the machine, is not a deficiency to be corrected. It is the source of our humanity.
Consider a six-word story apocryphally attributed to Ernest Hemingway:
For sale: baby shoes, never worn.
Seventy-two bits of information. And yet to understand it, to truly understand it, to feel what it means, you need to have lived. You need to carry within you a knowledge of what it is to hope for a child, and to lose one. Not as an abstraction, but as a possibility that belongs to your own life. The machine can reproduce those words. The machine cannot feel their weight, because feeling their weight requires being the kind of creature that is born, that suffers, and that dies.
What machines are missing, when it comes to the decisions that matter most, is two things. First, the accumulated cultural understanding we have built across generations — the shared context through which we interpret the world, transmitted through that narrow bandwidth channel across millennia. Second, and more profoundly, exposure to human vulnerability. We are accountable to one another because we share the same condition. We can be hurt in the same ways. This is not a weakness: it is the foundation of conscience, of compassion, of what we recognise as human dignity.
Human Analogue Machines
The large language models, the systems like ChatGPT that have so captivated and alarmed the world, are not (as they are sometimes presented) alien intelligences arriving from outside our culture. I prefer to call them human analogue machines. They were trained on everything we have written: every novel, every scientific paper, every letter, every prayer, every argument. They have distilled something of our cultural expression. But what they have absorbed is the pattern of our thinking, not its source. They have the shadow, not the light.
And here is the risk: a sufficiently sophisticated shadow can be mistaken for the real thing. The algal bloom looks, from a distance, like life. It is, in one sense, life. But it is a monoculture that consumes the oxygen and crowds out everything else.
Human Culture as Ecology
Human culture is an ecology. It has evolved across tens of thousands of years. It carries complexity that we cannot fully articulate because that complexity lives in relationships, in communities, in traditions, in liturgy, in stories passed between generations, in all the things that cannot be uploaded or scaled. The Church, more than almost any institution I can think of, understands this. It has been a custodian of this ecological complexity across centuries. It has preserved voices and faces through the long and turbulent ages of human change.
The information revolution is fast. Our ecology is slow. One of the most important things I argue in my book is that we are psychologically incapable of properly representing things that operate at radically different timescales to our own. Things that are vastly faster than us, or vastly slower. We cannot think on geological time, which is why we have struggled to grasp the damage we are doing to our natural environment. And we cannot think on machine time, which is why we struggle to grasp what the current information environment is doing to our culture.
The question before us is whether we can bring custodial wisdom to this new moment. Can we ensure that what floods in — the nitrogen of our digital age — feeds genuine growth rather than a catastrophic bloom?
I believe the answer is yes. But only if we are clear about what we are protecting and why.
The Atomic Human
We are protecting the space for what I call the atomic human. The idea comes from Democritus, who argued that if you keep cutting matter in two, eventually you reach a piece that cannot be cut further, an indivisible atom. The same question applies to human capability: as machines slice away what we once thought was uniquely ours — calculation, chess, diagnosis, composition — is there an indivisible kernel that remains? I believe there is. And that kernel is precisely what this conference seeks to preserve.
Our voices and our faces are worth preserving because they carry what cannot be transferred: the knowledge that comes from being mortal, from being vulnerable, from being, in the most profound sense, one of us.
Thanks!
For more information on these subjects and more you might want to check the following resources.
- company: Trent AI
- book: The Atomic Human
- twitter: @lawrennd
- podcast: The Talking Machines
- newspaper: Guardian Profile Page
- blog: http://inverseprobability.com