Why We Count the Risks Before We Can See the Benefits
What If Our Technology Ledger Is Incomplete?
Whenever a powerful new technology appears, we begin keeping a ledger.
On one side, we write the risks: jobs that may disappear, skills people may lose, new forms of crime, new opportunities for manipulation, new dependencies, new inequalities and new accidents waiting to happen.
And we should. Powerful technologies deserve serious scrutiny precisely because they are powerful.
But there is something peculiar about the other side of the ledger. It is much harder to fill in.
How do you count the benefit of an application that nobody has invented yet? How do you count the jobs that do not yet have names? How do you measure the industries that may emerge to solve problems the technology itself creates? How do you calculate the consequences of not adopting a useful technology?
How do you count a disaster that never happened because a new system prevented it? And when a technological benefit eventually becomes ordinary life, do we even continue recognising it as a benefit of the technology?
These questions became the foundation of The Asymmetric Technology Ledger.
This Is Not an Argument That AI Is Harmless
This series is not an argument that artificial intelligence is harmless. It isn't.
AI can be misused. It can make mistakes. It can displace workers. It can concentrate power. It can facilitate fraud, manipulation and surveillance. Poorly designed systems can cause real harm.
Nor is this series an argument that because humanity survived previous technological revolutions, everything will automatically work itself out this time. History gives us no such guarantee.
The argument is different.
Identifying a risk is not the same thing as completing a risk-benefit analysis.
The Asymmetry of What We Can See
We are remarkably good at imagining what a new technology might destroy because the things at risk already exist in front of us.
The job exists. The profession exists. The institution exists. The current way of doing something exists.
So when technology threatens any of them, the potential loss is visible.
But many potential gains lie on the other side of invention. Before the aircraft, there was no global aviation system whose benefits could be entered into a forecast. Before the internet, nobody could comprehensively list the businesses, professions, relationships and services that would eventually depend upon it.
Before widespread computing, many of today's occupations could not have appeared in an employment forecast because the problems they solve, and sometimes even the words used to describe them, did not yet exist.
That creates an asymmetry. We compare visible things that might be lost with invisible things that might eventually be created.
New Problems Create New Responses
There is another side we frequently forget: problems themselves create responses.
Aircraft can crash, so societies developed aviation safety engineering, air-traffic control, accident investigation, maintenance systems and regulation.
The internet created cybercrime, and cybercrime helped create an enormous cybersecurity industry.
If AI creates new dangers, society will need people capable of understanding, detecting, regulating, investigating and reducing those dangers. If AI disrupts employment, societies will have to respond to that disruption.
If automation changes how income circulates through an economy, businesses and governments may eventually have to reconsider how purchasing power reaches the people expected to consume what increasingly automated systems produce.
None of those responses guarantees that AI will create more jobs than it removes. None proves that its benefits will outweigh its harms. That would simply replace pessimistic certainty with optimistic certainty.
Uncertainty Requires Serious Inquiry
The more interesting position is uncertainty accompanied by serious inquiry.
Throughout this series, questions from both sides of the ledger are being asked.
- What could AI take away, and what might it make possible?
- What dangers does adoption create, and what are the consequences of non-adoption?
- Which existing jobs might disappear, and what new work becomes necessary because AI exists?
- What happens when technology fails, and what happens when it succeeds so consistently that we stop noticing the benefit?
And perhaps most importantly:
Are we evaluating the future using only the categories available to us in the present?
What the Series Will Explore
The articles that follow move from writing, printing, cars, aircraft, television and the internet into artificial intelligence, employment, regulation, economic transition and the professions and institutions that may not yet exist.
They are not intended to tell us that everything will be fine. They are an invitation to think more completely about what “everything” actually includes.
Because when the next transformative technology arrives, asking “What could go wrong?” is responsible. But treating the answers as the entire ledger is not.
There is another question worth asking alongside it:
“What could go right?”
The difficulty is that some of the most important answers may not have been invented yet.
Next: Part I
We Have Been Afraid of Technology Before
From Writing to Artificial Intelligence, Humanity Has Had This Conversation More Than Once.