Why Successful Technology Eventually Disappears From the Ledger
Something interesting happens when technology works.
We stop noticing it.
Send an email.
It arrives.
Nothing happened.
Open a map on your phone.
A satellite system calculates your location and guides you through a city you have never visited.
Nothing happened.
Transfer money electronically.
It arrives almost instantly.
Nothing happened.
Video-call somebody thousands of miles away.
You see their face in real time.
Nothing happened.
Search decades of accumulated human knowledge in seconds.
Nothing happened.
These things are so ordinary that describing them as technological miracles would sound melodramatic.
But now let something fail.
A hacker steals money.
Cybercrime.
A social-media algorithm spreads misinformation.
Technology harms democracy.
A navigation system directs somebody incorrectly.
GPS failure.
An AI model invents information.
AI hallucination.
Suddenly the technology becomes visible again.
This creates a strange accounting problem.
Technological success gradually becomes ordinary life.
Technological failure remains technological failure.
Imagine newspapers reporting aviation using the opposite convention.
“ANOTHER 100,000 FLIGHTS LAND SAFELY.”
Every day.
Nobody would read it.
A plane crash, correctly, becomes international news.
But psychologically this means extraordinary reliability can become invisible while exceptional failure dominates our perception of the system.
Something similar happens with technological benefits across generations.
There were once people who lived without instantaneous long-distance communication.
Without automobiles.
Without aviation.
Without television.
Without antibiotics.
Without search engines.
Without video calls.
Without online banking.
Without satellite navigation.
The generations born afterwards don't experience these things primarily as technological benefits.
They experience them as the minimum expected conditions of life.
We inherit yesterday's technological revolution as today's baseline.
Then a new technology appears.
Its benefits are uncertain because many haven't been invented yet.
Its risks are unfamiliar and therefore highly visible.
We compare:
the frightening possibilities of tomorrow's technology
against
a present whose technological miracles we have forgotten are technological miracles.
That is not a fair comparison.
Imagine somebody in 1995 evaluating the internet.
They can already imagine pornography.
Fraud.
Hacking.
Extremist material.
Strangers contacting children.
Privacy problems.
But ask them to put the benefits on the other side.
Could they confidently write:
- global cloud computing;
- modern open-source software collaboration;
- high-definition international video calls;
- mass remote working;
- instantaneous global banking;
- streaming education;
- modern e-commerce;
- collaborative scientific databases;
- smartphone navigation integrated with live traffic?
Some might imagine pieces.
Nobody could accurately enumerate the world that would emerge.
The risks existed in the present imagination.
Many benefits existed only in the future creativity of millions of people who had not yet built them.
And now we stand in that same uncomfortable position with artificial intelligence.
We can see hallucinations.
We can see deepfakes.
We can see cheating.
We can see job displacement.
We can see fraud.
We should respond to them.
But somewhere today somebody may be developing an AI application whose eventual importance we cannot yet understand.
Perhaps in medicine.
Perhaps materials science.
Perhaps accessibility.
Perhaps energy.
Perhaps education.
Perhaps engineering.
Perhaps something for which we do not yet have a category.
That possibility doesn't cancel the risks.
It completes the ledger.
And there is one final irony.
We are discussing whether artificial intelligence is too dangerous...
using computers...
connected through the internet...
displaying information on screens...
communicating through technologies that previous generations themselves regarded with varying degrees of suspicion.
Yesterday's controversial technology has become the platform from which we debate today's controversial technology.
Perhaps AI will eventually join them.
Perhaps some of today's fears will prove embarrassingly exaggerated.
Perhaps others will prove frighteningly accurate.
Most likely, if history is any guide, reality will contain both.
The challenge is therefore not to choose between technological worship and technological fear.
It is to become better at something humanity has repeatedly struggled to do:
Hold risk and possibility in our minds at the same time.
Take danger seriously without turning possibility into inevitability.
Protect people without protecting them from progress itself.
Learn the technology well enough to govern it.
Help those genuinely harmed by the transition.
Build safeguards around capabilities worth keeping.
And remember that refusing an opportunity also has consequences.
Perhaps the question we should carry into every technological revolution is therefore not:
“What could go wrong?”
We absolutely should ask that.
But when the room falls silent after the answers, somebody should ask the other question too:
“What could go right?”
History suggests we are considerably worse at predicting the answer to that one.
Read the series introduction: The Asymmetric Technology Ledger →