AI, Money and the Value of Knowledge
Artificial intelligence is often described as one of the greatest technological advances of our age. It can write software, analyse medical information, summarise huge quantities of text, translate languages, generate images, assist engineers and help ordinary people solve problems that would once have required specialist knowledge.
But there is an uncomfortable question sitting behind all of this progress.
“What happens when the machines become so hungry for knowledge that we start destroying the knowledge itself in order to feed them?”
That might sound dramatic, but it points toward a much larger problem in modern society. It is not really a problem about AI.
It is a problem about values.
The strange hunger of artificial intelligence
Modern artificial intelligence requires enormous quantities of information.
Books, websites, scientific papers, manuals, conversations, computer code, photographs, historical records and almost every other form of human-created information can potentially contribute to the development of AI systems.
That makes high, quality human, created material extremely valuable.
Older books are particularly interesting because they come from a world before generative AI.
They were written entirely by people.
They contain human opinions, mistakes, discoveries, observations, techniques, experiences and ways of thinking that originated independently of artificial intelligence.
That distinction may become increasingly important.
The internet is already beginning to contain huge quantities of AI-generated material. As that proportion increases, future AI developers face a peculiar problem.
They risk training artificial intelligence on information that was itself generated by artificial intelligence.
Eventually the snake begins eating its own tail.
Errors can be repeated.
Ideas can become homogenised.
Original sources can disappear beneath thousands of rewritten versions.
The information may still look convincing, but its connection to the original human observation becomes weaker.
That makes old human-created material enormously valuable.
And this is where things start becoming absurd.
A book is more than training data
A physical book is an extraordinary piece of technology.
It does not require electricity.
It does not require an internet connection.
It does not need a subscription.
It cannot suddenly disappear because a company changes its terms of service.
It cannot be remotely altered.
It can survive for centuries.
A person can take it from a shelf, open it and read exactly what the author wrote.
That makes books exceptionally robust stores of knowledge.
Artificial intelligence works very differently.
An AI model trained on a thousand books does not become a library containing those thousand books.
It learns statistical relationships from the information.
It may understand concepts found within those books. It may answer questions based partly on patterns learned from them.
But it does not preserve those books in their original form.
This distinction is crucial.
Destroying a book after using it to train an AI system is rather like destroying the engineering drawings after teaching an apprentice how to build the machine.
The apprentice may have learned a great deal.
But the apprentice is not the drawing.
Details will be forgotten.
Some things will be misunderstood.
Other things may never have been absorbed at all.
The original source still has value.
Perhaps greater value than before.
Digitisation and destruction are not the same thing
There is nothing inherently wrong with digitising books.
In fact, digitisation can be one of the greatest tools available for preserving knowledge.
A fragile historical book can be scanned and copied around the world.
Researchers can search it instantly.
People who would never have been able to visit the physical archive can read it from home.
Digital preservation and physical preservation can work beautifully together.
But digitisation does not require destruction.
There may be circumstances where industrial scanning methods make destructive processing (destructive scanning) cheaper or faster, particularly when dealing with enormous quantities of material.
And there lies the problem.
The cheapest solution economically may not be the best solution culturally.
If the decision is reduced entirely to cost, a simple calculation appears:
Destroying the book costs less.
Therefore destroy the book.
The spreadsheet has spoken.
But the spreadsheet cannot calculate historical importance.
It cannot know whether an obscure technical manual sitting in a second-hand bookshop contains information that will become valuable fifty years from now.
It cannot calculate the emotional value of a first edition passed between generations.
It cannot predict which forgotten craft, engineering technique or historical observation future researchers may desperately want to rediscover.
Money can calculate price.
It cannot calculate importance.
When money becomes the scoreboard
Money is useful.
Anybody who has experienced life without enough of it knows exactly how important it becomes when bills need paying and food needs buying.
The problem is not that money exists.
The problem comes when money becomes the measurement system for everything else.
Modern society frequently asks:
How much is it worth?
How much profit can it generate?
How quickly can it be monetised?
How cheaply can it be processed?
Those are reasonable questions inside a business.
They become dangerous when they become the only questions civilisation asks.
Some things have value that cannot easily be expressed financially.
Knowledge.
History.
Natural ecosystems.
Community.
Human dignity.
Scientific understanding.
Cultural memory.
A society can become extremely efficient while simultaneously becoming extremely foolish if it optimises everything around financial return.
That is one of the great contradictions of technological progress.
We can become brilliant at doing things without becoming equally good at deciding whether we should do them.
Why doesn’t somebody simply stop it?
This raises another question.
Where are governments?
Governments regulate buildings, medicines, food production, financial markets, transport systems, telecommunications and thousands of other areas.
So why should preservation of knowledge receive less attention?
Part of the answer may be that the problem does not fit neatly into traditional regulation.
Governments understand copyright.
They understand ownership.
They understand taxation.
They understand competition law.
They understand national security.
But the question of whether commercially valuable historical knowledge should be physically preserved for future generations falls awkwardly between those categories.
A company may legally own a book.
If it legally owns that book, conventional thinking says it can do whatever it wants with it.
Usually that makes sense.
But what happens when millions of individually insignificant decisions collectively begin removing culturally important material from circulation?
At that point private ownership and public interest begin to overlap.
We already accept this principle in other areas.
Historic buildings may be protected.
Archaeological discoveries may be protected.
Endangered species may be protected.
Important artworks may be prevented from leaving certain countries.
National archives preserve government documents because society recognises that some things matter beyond their immediate financial value.
Why should knowledge be different?
A library for humanity
Perhaps the answer is not to stop artificial intelligence from learning from books.
Quite the opposite.
AI could become one of the greatest tools ever created for unlocking humanity’s accumulated knowledge.
But preservation should be built into the process.
Imagine a simple principle:
No unique or scarce source of human knowledge should be destroyed solely for the convenience of data processing.
If a book must be scanned, scan it.
If the binding must genuinely be removed, preserve the pages.
Create a verified digital archive.
Record the edition.
Record where it came from.
Preserve multiple copies in independent repositories.
And where possible, return physical books to circulation.
Artificial intelligence and libraries should be partners, not predator and prey.
Beyond money
This eventually leads to a much larger question.
Could civilisation operate differently?
Could we create a world in which money remains useful without allowing financial value to override every other kind of value?
A completely moneyless society would be extraordinarily difficult to design.
Money performs several important functions.
It helps distribute scarce resources.
It provides a common unit for exchange.
It rewards labour.
It allows millions of people who never meet one another to cooperate economically.
Simply removing money would not remove scarcity, power or inequality.
Another system would quickly take its place.
Credits.
Favours.
Influence.
Political authority.
Access.
Status.
The more realistic challenge may therefore be not eliminating money, but limiting what money is allowed to decide.
Certain things could be recognised as belonging to a category of shared human inheritance.
Knowledge archives.
Scientific records.
Historical documents.
Genetic seed banks.
Important cultural works.
Essential environmental resources.
Once something enters that category, profit would no longer be the final authority.
Commercial use might still be permitted.
But preservation would come first.
That would represent a subtle but profound change in priorities.
Profit would remain part of civilisation.
It simply would not be allowed to eat civilisation.
AI forces us to decide what we value
Artificial intelligence is going to force humanity to answer questions that technology previously allowed us to avoid.
What knowledge should be preserved?
Who owns information?
What happens when machines generate more text than humans?
How do we distinguish original observation from endlessly recycled synthetic material?
And perhaps most importantly:
What is something worth when its value cannot be measured in money?
These questions are bigger than artificial intelligence.
AI is simply shining a very bright light on them.
We are building machines capable of processing extraordinary amounts of human knowledge.
That should be an opportunity to preserve, organise and understand that knowledge better than any civilisation before us.
It should not require sacrificing the sources from which that knowledge came.
Progress should leave humanity with more understanding, not fewer originals.
Because once something truly disappears, neither money nor artificial intelligence can necessarily bring it back.
And perhaps that is the principle we should remember whenever technology presents us with another supposedly efficient shortcut:
“Before asking what something is worth, ask whether the world will be poorer when it is gone.“
Copyright© 2012-2026 Radionics Software. All rights reserved.