Imagine a country where elections are still held, parliament still exists, and citizens remain free to express their opinions — but where politics no longer works quite the way we are accustomed to.
A politician may promise lower taxes, higher benefits, greater defence spending, and a smaller national debt at the same time.
But the promise does not automatically become policy.
First, it passes through an independent economic council. Then through models estimating the consequences over five, twenty, and fifty years. The assumptions are made public. The opportunity costs are disclosed. If the numbers do not add up, the politician must either change the proposal or openly explain which other objective is being sacrificed.
Politics may still decide where society should go.
But it has far less freedom to pretend that natural laws, mathematics, and economic relationships cease to exist merely because an election is approaching.
We might call this Plato’s Republic 2.0.
The Philosopher King in a Suit
Around 2,400 years ago, Plato advanced an idea that remains provocative today. He doubted that political questions automatically produced wise outcomes simply because large numbers of people participated in deciding them.
In The Republic, he instead imagined a society governed by people who combined knowledge with moral character. The philosopher king would not seek power for personal gain, but would govern because he understood what a good society required.
A modern equivalent would probably not wear a toga and debate metaphysics in the town square.
She would be more likely to be an economist, engineer, physician, lawyer, physicist, or experienced civil servant.
But a problem immediately appears.
Knowledge is not enough.
A brilliant economist may be corrupt.
A distinguished physician may be narcissistic.
A professor may use extraordinary intelligence to rationalise whatever benefits his own social group.
A bureaucrat may want his department to grow.
A politician may understand perfectly well which policy is sensible in the long term and still choose something else because the next election is eighteen months away.
Plato’s real requirement for the ruler was therefore considerably more demanding than academic credentials.
Those who govern must both understand and be capable of placing their own interests last.
That is where the modern republic would have to begin.
Citizens Choose the Goals
In this hypothetical society, democracy is not abolished.
Instead, a much clearer distinction is made between two fundamentally different political questions.
The first is:
What do we want?
The second is:
How do we achieve it?
No expert can answer the first question on behalf of society.
How much economic equality do people want?
How much of their income should remain private?
How much security are they prepared to purchase at the cost of personal freedom?
How much consumption today should be sacrificed for investment that may not pay off for thirty years?
These are questions of values.
Here, the citizen remains sovereign.
In elections, the population therefore continues to determine society’s fundamental direction. Political parties compete over goals, priorities, and values.
But once the direction has been chosen, the system changes.
Politicians no longer enjoy unlimited authority to determine which empirical relationships are true.
A government may say:
We want to reduce emissions by 50 per cent without reducing GDP per capita.
That is a legitimate political objective.
But the government may not then simply declare that a particular energy technology will provide a particular amount of electricity in 2040.
That is an empirical question.
At that point, engineers, economists, and energy-system specialists take over.
Democracy chooses the destination.
Competence designs the route.
A Fourth Branch of Government
The modern state is usually divided into legislative, executive, and judicial power.
In Plato’s Republic 2.0, a fourth branch would effectively be added:
epistemic power.
Its purpose would not be to tell citizens what to think.
Its purpose would be to determine what the state can reasonably claim to know.
An independent fiscal council would assess public finances.
An energy council would evaluate physical capacity and security of supply.
A demographic institute would model pensions, population, and labour supply.
Medical expert bodies would assess healthcare interventions.
Defence planning would be led by people who actually understand military logistics, weapons systems, and strategic endurance.
None of these groups would be omniscient.
On the contrary, the system would be designed around the assumption that experts can also be wrong.
Every major decision would therefore have to include its models, assumptions, uncertainty ranges, and competing analyses, all open to public scrutiny.
The expert’s authority would not come from the title.
It would come from the fact that the reasoning could be examined.
The Professor Is Not Automatically Right
This matters.
Otherwise, a Platonic society could quickly degenerate into something much less attractive: a permanent academic ruling class proclaiming that it alone represents reason.
But a university degree guarantees neither judgement nor intellectual honesty.
A professor of literature does not become an expert on public finance merely because the word “doctor” appears before the name.
An economist does not become an expert on nuclear physics.
A physicist does not automatically become qualified to organise a healthcare system.
And even two experts in the same field may arrive at entirely different conclusions.
The system is therefore not about academics ruling.
It is about political authority within a particular domain having to be justified by relevant competence.
An entrepreneur with thirty years of experience in industrial production might, in some matters, possess more relevant knowledge than someone who has spent an entire career at a university.
Knowledge must be tested against the problem.
Not against the business card.
The Problem of Self-Interest
But even if we somehow managed to identify the most competent people in the world, humanity’s oldest political problem would remain:
Power changes incentives.
The person administering a system soon acquires an interest in the system becoming larger.
The person controlling an agency may begin to confuse the interests of that agency with the interests of society.
The member of an expert class may begin to regard criticism of the expert class as criticism of knowledge itself.
The modern philosopher king would therefore have to be constrained far more tightly than most politicians are today.
Plato would probably recognise the logic immediately.
The most important decision-makers might, for example, face extensive requirements for financial transparency, strict conflict-of-interest rules, limitations on outside employment, and long cooling-off periods before being allowed to take jobs in companies or organisations affected by their previous decisions.
The purpose would not be to find people without self-interest.
Such people may not exist.
The purpose would be to build institutions in which self-interest becomes expensive, visible, and difficult to convert into power.
And Then Comes AI
At this point, the experiment becomes even more interesting.
For the first time in history, humanity may be able to construct a decision-support system that lacks at least some of the biological motivations that have always accompanied human rulers.
An AI does not need to be re-elected.
It does not need a higher salary.
It has no children who need board positions.
It does not become offended because a newspaper wrote something unpleasant about it.
It does not need to win the next internal power struggle at the Ministry of Finance.
One could therefore imagine AI taking the role of the state’s great consequence machine.
Every major reform would be run through transparent models.
What happens to growth?
Employment?
Public debt?
Energy supply?
The housing market?
The pension system?
What happens after two years — and after thirty?
The AI could also be instructed to actively attempt to falsify the government’s own assumptions.
It would function less as a ruler and more as an institutional devil’s advocate.
But AI does not solve Plato’s problem completely.
Even the most advanced model must still be told what to optimise.
Imagine two reforms.
One produces higher economic growth but greater income inequality.
The other produces lower growth but a more equal distribution of income.
Which one is better?
There is no equation capable of answering that question on its own.
Someone must decide how those outcomes are valued.
And suddenly we are back where we started:
with the citizen.
Election Day in Republic 2.0
What would an election look like?
Perhaps much like today.
Political parties would present their visions.
One party wants a smaller state.
Another wants a larger public sector.
A third prioritises climate policy.
A fourth prioritises defence.
A fifth places greater emphasis on economic equality.
But next to every campaign promise would appear something that political advertising rarely contains today:
the price tag.
And not merely in monetary terms.
If the state spends one hundred billion here, voters must be shown which alternative uses of those resources disappear as a consequence.
If the retirement age is not increased despite rising life expectancy, the consequences for future taxes or pensions must be disclosed.
If an industry is subsidised, the estimated cost per job created must be shown.
If a tax cut is proposed, its long-term financing must be explained.
The politician would still be free to say:
I believe this is worth the cost.
But it would become much harder to say:
There is no cost.
That is a small distinction.
But potentially a revolutionary one.
Democracy’s New Division of Labour
The result would be neither pure democracy nor pure technocracy.
It would be a division of labour.
The people choose the values.
Politics formulates the goals.
Experts assess the methods.
AI analyses the consequences.
Independent institutions scrutinise all four.
There are, of course, serious objections to such a system.
An expert-driven model can select people from similar backgrounds who share similar blind spots.
What appears to be neutral expertise can therefore acquire a social, ideological, or institutional bias.
That is why opposition, free media, open data, and the ability to remove experts from power would remain essential.
No one may become infallible.
Not the voter.
Not the professor.
Not the bureaucrat.
And certainly not the AI.
Plato’s Old Question
Perhaps that is what makes Plato’s thought so uncomfortably relevant after more than two millennia.
We accept without much argument that a person requires years of training before being allowed to operate on a heart.
We demand certification before someone can fly a passenger aircraft.
We expect the engineer designing a bridge to understand structural mechanics.
Yet some of the most complex systems human beings have ever created — public finances, energy systems, pensions, international security — are ultimately governed through political processes in which the ability to persuade the electorate can sometimes matter more than the ability to understand the system itself.
Plato’s answer was the philosopher king.
Our answer need not be the same.
Perhaps the modern solution is not to find the perfect ruler.
Perhaps it is to build a system in which no single person needs to be perfect.
A society in which democracy still decides what is desirable, but reality is given a veto over what is possible.
Where knowledge grants influence, but never immunity from scrutiny.
Where power is separated from personal gain as far as institutions can reasonably achieve.
And where AI does not become our new king, but instead the merciless adviser sitting quietly at the back of the room who, every time someone promises everything to everyone, raises one simple question:
“And how exactly were you planning to pay for that?”