From the earliest formation of the brain in utero to the final moments before death, we are learning organisms. That much has not changed. What has changed—profoundly—is the environment in which that learning takes place.
In 2013, I understood this environment as “the Web”: a vast, distributed archive of human knowledge, increasingly accessible to anyone with a device and connection. It democratised access, expanded opportunity, and blurred the boundaries between formal and informal education. The global village became a kind of digital fireplace—people gathering around shared knowledge, ideas, and networks. I wrote about it here.
That still holds. But it is no longer sufficient.
We are now moving from access to information towards interaction with intelligence. The Web is no longer simply something we consult. It is becoming something that responds, adapts, and, crucially, remembers.
The interface has shifted—from person and information to person and a system that learns about that person.
Nowhere is this more significant than in medicine.
A decade ago, I was already circling a problem that remains stubbornly unsolved: information alone does not change behaviour. A patient prescribed a preventer inhaler for asthma may fully understand its purpose and still not take it. The reasons are rarely informational. They are behavioural, emotional, and bound up in identity: denial, resistance, forgetfulness, inconvenience, or fatigue in response to a chronic condition.
The Web, as I saw it then, could inform. But it could not accompany.
Recently, during a period when my asthma deteriorated markedly—worse than at any point in forty years—I found myself relying on AI in precisely this companion role. Not for diagnosis, but for something arguably more important: continuity. It helped me interpret fluctuating peak flow readings, weigh decisions about whether to coach or rest, and articulate clearly—both to myself and to others—the nature of what I was experiencing. It allowed me to think through exposure to triggers, the cumulative impact of long poolside hours, and the need to adapt my working pattern without guilt or guesswork.
In short, it sat alongside me—not replacing clinical advice, but helping me make sense of it, and act on it.
This is the shift.
What is now emerging is something closer to what I once described as an “artificial companion”—a system that sits, metaphorically, on the shoulder. Not as authority, but as presence. Not simply instructing, but engaging.
Such a system explains complex medical knowledge in a personally meaningful way. It supports behaviour—nudging, reminding, and adapting to routines. And it engages emotionally—recognising hesitation, resistance, or drift, and responding accordingly.
This reshapes the doctor–patient relationship. It is no longer a simple exchange between clinician and individual, but a triad: patient, clinician, and a continuous, responsive system that exists between appointments—where most health decisions are actually made.
Traditional medicine is episodic. Chronic conditions are not.
More broadly, this reframes learning itself. It is no longer just about acquiring knowledge, but about regulating behaviour over time—remembering, deciding, persisting. The learner is not simply a sponge or participant, but a system navigating another system.
At the far end of life, another question emerges. If aspects of our decisions, language, and patterns are captured and modelled, what remains when we are gone?
In 2013, I asked what role the Web might play across the span of a life. The answer now is clearer.
We are moving from a world in which we access knowledge, to one in which knowledge systems accompany us—shaping not just what we understand, but how we live.
From Web to Companion: How Learning—and Medicine—Has Changed
From the earliest formation of the brain in utero to the final moments before death, we are learning organisms. That much has not changed. What has changed—profoundly—is the environment in which that learning takes place.
In 2013, I understood this environment as “the Web”: a vast, distributed archive of human knowledge, increasingly accessible to anyone with a device and connection. It democratised access, expanded opportunity, and blurred the boundaries between formal and informal education. The global village became a kind of digital fireplace—people gathering around shared knowledge, ideas, and networks.
That still holds. But it is no longer sufficient.
We are now moving from access to information towards interaction with intelligence. The Web is no longer simply something we consult. It is becoming something that responds, adapts, and, crucially, remembers.
The interface has shifted—from person and information to person and a system that learns about that person.
Nowhere is this more significant than in medicine.
A decade ago, I was already circling a problem that remains stubbornly unsolved: information alone does not change behaviour. A patient prescribed a preventer inhaler for asthma may fully understand its purpose and still not take it. The reasons are rarely informational. They are behavioural, emotional, and bound up in identity: denial, resistance, forgetfulness, inconvenience, or fatigue in response to a chronic condition.
The Web, as I saw it then, could inform. But it could not accompany.
Recently, during a period when my asthma deteriorated markedly—worse than at any point in forty years—I found myself relying on AI in precisely this companion role. Not for diagnosis, but for something arguably more important: continuity. It helped me interpret fluctuating peak flow readings, weigh decisions about whether to coach or rest, and articulate clearly—both to myself and to others—the nature of what I was experiencing. It allowed me to think through exposure to triggers, the cumulative impact of long poolside hours, and the need to adapt my working pattern without guilt or guesswork.
In short, it sat alongside me—not replacing clinical advice, but helping me make sense of it, and act on it.
This is the shift.
What is now emerging is something closer to what I once described as an “artificial companion”—a system that sits, metaphorically, on the shoulder. Not as authority, but as presence. Not simply instructing, but engaging.
Such a system explains complex medical knowledge in a personally meaningful way. It supports behaviour—nudging, reminding, and adapting to routines. And it engages emotionally—recognising hesitation, resistance, or drift, and responding accordingly.
This reshapes the doctor–patient relationship. It is no longer a simple exchange between clinician and individual, but a triad: patient, clinician, and a continuous, responsive system that exists between appointments—where most health decisions are actually made.
Traditional medicine is episodic. Chronic conditions are not.
More broadly, this reframes learning itself. It is no longer just about acquiring knowledge, but about regulating behaviour over time—remembering, deciding, persisting. The learner is not simply a sponge or participant, but a system navigating another system.
At the far end of life, another question emerges. If aspects of our decisions, language, and patterns are captured and modelled, what remains when we are gone?
In 2013, I asked what role the Web might play across the span of a life. The answer now is clearer.
We are moving from a world in which we access knowledge, to one in which knowledge systems accompany us—shaping not just what we understand, but how we live.
