I anticipate that this article will at some point be overtaken by historical events. If I have read the tea leaves at the bottom of my china cup correctly, a “mother of all societal cleanups” is coming, exact timing TBD. My reading of events is that the high end of global networks of cartels and traffickers is already being addressed, quietly and mostly out of public sight. That still leaves a vast body of compromised apparatchiks and low-level grifters who have been bleeding our world dry.
I believe justice is coming — at scale.
If I am wrong about the scale or timing of that cleanup, the argument that follows still stands, only stronger.
A cleanup is not enough
Even the most comprehensive cleanup would leave a residue of corruption from which the old patterns could regrow. Move every transaction onto a transparent and immutable blockchain; rigorously enforce the law on identity and accounting procedure; punish breaches of trust and duties of care—and still the self-serving will seek advantage over everyone else.
There will never be a final victory in which corruption simply disappears. Society therefore needs a permanent immune system.
That means distributing the means of accountability. Ordinary people need the practical ability to preserve evidence, reconcile records, expose contradictions, trace transactions, challenge false institutional narratives, and place defensible findings into the public domain.
The answer to permanent human fallibility is permanent auditability.
We cannot abolish the self-serving. We can make predation harder to conceal, collusion harder to sustain, fraud cheaper to expose, and institutional lies more dangerous to tell. Corruption can only be contained by constant vigilance—and increasingly, that vigilance will be digitally armed.
By “digital arms” I mean the lawful tools of evidential self-defence: recording, preservation, search, timeline reconstruction, financial reconciliation, contradiction detection, adversarial analysis, and publication. AI radically reduces the expertise and labour previously needed to wield them.
What once required a law firm, investigative newsroom, regulator, or intelligence unit can increasingly be attempted by an ordinary person—for good or evil purposes.
Hence audit technology cannot be the final answer, however important a part it has to play.
The wounded auditor
The problem is not merely one of databases, evidence and clever technology for spotting discrepancies. I have lived through multiple institutional malfunctions. My whole “ghost court” escapade in Cumbria unfolded against the backdrop of a horrendous concurrent family situation that I was helping to manage in America.
The trauma of one fed into the other.
In that context, the most basic forensic task—assembling documents into a timeline—can itself require you to push through unresolved trauma. Open the letter. Revisit the event. Find the earlier email. Read the allegation again. Compare it with the order. Work out what happened next. Repeat.
Half that battle is overcoming the “freeze-up” of a nervous system that has learnt to experience the material itself as a threat.
Victims of institutional wrongdoing can therefore be at their weakest precisely when the system demands that they become their own investigators, archivists, advocates and project managers. That problem is especially acute in the social fallout from the Covid era, when many people emerged with damaged finances, fractured relationships, disrupted lives and profoundly diminished trust in public institutions.
This is where AI offers something qualitatively new. It is not merely an investigative tool for finding discrepancies faster. It can become a cognitive exoskeleton for the wounded auditor, taking on some of the burden of remembering, organising, comparing and revisiting material that the human being may find painful even to confront.
I have reached the point where I sigh with relief whenever I look in my mailbox and find it empty.
That is not a healthy relationship between citizen and state.
⁂
I increasingly refuse to participate voluntarily in administrative schemes that I regard as predatory or illegitimate.
Clean-air charging regimes are one example: where I see environmental justification being used to normalise surveillance, restriction and revenue collection around ordinary movement, my instinct is no longer to presume benign administration and comply unthinkingly. It is to demand that the claimed authority, necessity, proportionality and use of the resulting powers be demonstrated.
Others will draw those boundaries differently. That is not the essential point.
The essential point is what happens when institutional trust has been exhausted. Every brown envelope acquires the accumulated weight of the ones before it. A demand that might once have been treated as routine administration is instead received as a potential threat requiring investigation.
The cognitive cost of citizenship rises.
Statutory authority answers whether the state may act. It does not, by itself, answer whether a particular exercise of that authority was necessary, proportionate, honest, competent, or directed towards its proper purpose.
This is why institutional legitimacy cannot consist merely in possessing statutory powers. A state demanding obedience while insulating its own failures from meaningful scrutiny consumes the trust on which inexpensive government depends. Eventually compliance has to be obtained through ever more monitoring, penalties, enforcement and compulsion because voluntary legitimacy has been spent.
None of this requires believing that the state itself is illegitimate.
I do not. Courts, public administration, taxation and collective infrastructure are necessary features of a complex society. The distinction I want to make is between the legitimate functions of the state and criminality that can become embedded within, adjacent to, or protected by its machinery.
If such a criminal wing exists, the answer is not to destroy the organism.
It is to identify the diseased tissue, establish what actually happened, hold individuals to account under due process, repair the institutional mechanisms that protected them, and pare the corruption away.
And then comes the harder part: stopping it from growing back.
Accountability is a two-way street
Here comes the catch. The public are not necessarily any more virtuous than their public servants. Indeed, often quite the reverse.
An honest family-law judge can diligently weigh contradictory evidence in an appallingly difficult case, make the best decision available, and still be accused of bias by the losing party.
A traffic officer can conduct a lawful stop and subsequently encounter a narcissistic and entitled driver determined to manufacture misconduct.
A genuine criminal can writhe and twist through every available procedural opening, generating allegations, appeals and complaints whose purpose is not accountability but delay and evasion.
Giving such people AI does not make them honest. It makes them dangerous.
“Digital arms” therefore cannot mean democratising the ability to manufacture plausible accusations faster than institutions can answer them. That would simply create another form of asymmetric warfare—and eventually make conscientious public service impossible.
The public must be audited too.
What we need is something closer to a meta-audit of the relationship between citizen and state:
What may we reasonably demand from an institution?
What constitutes evidence of malfunction rather than merely an outcome we dislike?
Which errors indicate negligence or corruption, and which belong to the unavoidable residue of human judgement?
At what point does persistent scrutiny become harassment?
Conversely, at what point does institutional resistance to scrutiny become evidence in its own right?
Systems engineering offers a useful concept here: the predictable region of operation.
No serious engineer demands that a system perform perfectly under every imaginable combination of circumstances. We establish an operating envelope within which specified behaviour can reasonably be expected, identify known failure modes, build mechanisms for detecting excursions, and provide ways to recover when things go wrong.
The administrative state deserves the same intellectual fairness.
Courts cannot guarantee that every factual finding is correct. Police cannot prevent every crime. Doctors cannot save every patient. Regulators cannot anticipate every fraud.
Human beings will misunderstand one another, witnesses will remember events differently, documents will go missing, professionals will make mistakes, and occasionally several individually reasonable decisions will combine to produce a disastrous result.
None of that, by itself, establishes corruption.
The proper question is whether the system remained within its legitimate and predictable operating envelope:
Were the rules knowable?
Was relevant evidence considered?
Were powers exercised for their stated purposes?
Were conflicts disclosed?
Can the decision be reconstructed afterwards?
Was meaningful correction available?
Did those responsible act honestly within the discretion entrusted to them?
And the same disciplines must apply to the citizen making the accusation:
Did we preserve contrary evidence as carefully as favourable evidence?
Did we distinguish what we know from what we infer?
Have we tried to falsify our own theory?
Are we demanding a standard of institutional perfection that no human organisation could satisfy?
Would we accept the same reasoning if it supported the other side?
These are not concessions to institutional power. They are what make public audit credible.
Accountability is a two-way street.
The state must not convert authority into impunity, while the citizen must not convert accountability into persecution. The purpose of digital arms is not to ensure that everyone can win every dispute with authority. It is to make departures from legitimate process easier to detect, demonstrate, correct and—where necessary—punish.
We are not engineering a world without human failure and misadventure.
We are engineering one in which honest failure can be distinguished, as reliably as possible, from abuse.
The unbalanced books
As its name suggests, accountability is ultimately a form of accountancy.
I do not mean that everything reduces to money. I mean that human affairs take place across multiple notional ledgers, upon which events are recorded and where irreconcilable entries eventually demand explanation.
Consider a divorce:
There is a cost-based ledger: income, debts, property, transfers and assets to distribute.
There is a contractual ledger in the broad sense, encompassing positive law: rights, duties, agreements, jurisdiction and the equitable relief available to the court.
Behind both sits the covenantal ledger that no accountant or judge can completely capture: who behaved honourably, who sacrificed, who deceived, who inflicted harm, and what would actually constitute a righteous outcome.
Justice does not require these ledgers to be identical. It does require that they not be forced grotesquely apart.
Indeed, some of the worst injustices can appear perfectly legal. That is part of what makes them so durable. Once an outcome has been clothed in statutory authority, contractual consent, procedural regularity or judicial approval, legality itself can become a shield against reopening the underlying moral account.
The contractual ledger says: permitted.
The covenantal ledger may still say: wrong.
Yet forcing them apart is exactly the game that dishonest people learn to play:
A technically permissible transaction can conceal an inequitable purpose.
A procedurally valid act can generate a substantively outrageous result.
A truthful statement can be selected so as to create a profoundly false impression.
Each individual entry can appear defensible while the consolidated accounts tell another story.
There is a reason this happens that does not require everyone involved to be wicked. Ordinary human cognition is limited.
“I was only following the rules” has a genuine justification, up to a point. Rules allow us to inherit compromises without reopening every argument that produced them. They make behaviour predictable. They reduce the cognitive cost of cooperation. They let strangers coordinate without renegotiating first principles every morning.
That is why even tolerably bad rules should ordinarily be followed. The alternative is not instant justice but a regression towards self-help, where everyone becomes legislator, judge and enforcement officer in their own cause.
But obedience cannot be absolute.
There comes a point at which procedural regularity cannot discharge moral responsibility for a grotesque violation of innate rights or participation in heinous wrongdoing. No rulebook can perform that moral accounting on our behalf.
There is an uncomfortable symmetry here too.
The administrative state does not become inhuman solely because administrators are inhuman. The public places impossible and contradictory demands upon it. People weaponise police, courts, complaints procedures, regulators and family law against one another. They lie. They exaggerate. They conceal evidence.
They demand intervention when it benefits them and denounce authority when it does not.
Functionaries encounter this behaviour every working day. They develop shortcuts and defensive routines. They become suspicious. Procedures harden. Discretion contracts. Colleagues protect one another. Eventually mechanisms developed partly to survive public dysfunction can themselves become mechanisms for institutional dysfunction.
Citizen and state can corrupt one another.
That observation has become personally uncomfortable for me. Since Covid I have pushed back hard against state authority. Looking backwards, I can identify places where I overreacted or demanded more than an imperfect human institution could reasonably provide. Those entries belong in my ledger. I do not get to erase them because I believe my larger concerns were justified.
Conversely, I can identify occasions on which I believe institutions did me serious wrong. Acknowledging my own mistakes does not cancel theirs either.
That is what accounting means.
The purpose is not to arrange every transaction so that our preferred party always emerges with a credit balance. It is to keep all the books, including the uncomfortable ones, and reconcile them as honestly as humanly possible.
A civilisation capable of doing that at scale would possess something considerably more valuable than transparency.
It would possess the machinery of repentance.
After the takedown
Let us indulge the optimists for a moment.
Suppose the system of debt slavery disappears. You no longer owe the state merely for existing on the land, nor surrender a large share of your labour to sustain an administrative machine. Usury and manufactured financial distress cease to be mechanisms for stripping people of their assets. The centralised state contracts to a tenth of its present footprint, with family, church, community, charity and private enterprise recovering functions accumulated by bureaucracy.
Suppose technology delivers too. AI and robotics abolish much compulsory drudgery. Cheap energy and extraordinary advances in medicine and nanotechnology create forms of abundance that today seem fantastical. Competition for basic resources declines, and with it much of the conflict presently requiring adjudication and state coercion.
Suppose, even, that the criminal-industrial complex is dismantled while most of us do little more heroic than hold the line as forces beyond our ken perform the heavy lifting.
Then what?
We are still left with an engineering problem.
Human affairs must somehow reconcile at least three different kinds of ledger:
There is the covenantal ledger: what is right; what we owe one another by virtue of our humanity and relationships; the realm for which natural law provides a useful proxy.
There is the contractual ledger: what we actually promised, which rules we agreed to live under, which powers we delegated to institutions, and where the boundaries of those agreements and delegations lie.
And there is the cost-based ledger: who contributed what, who consumed what, who damaged what, who received what, and who should ultimately bear the material burden.
These ledgers overlap, but they are not interchangeable:
Something can be contractually permissible and covenantally repugnant.
Something can be covenantally desirable while exceeding the legitimate authority of the institution being asked to provide it.
A financial balance can be arithmetically impeccable while arising from an exploitative transaction.
A person can suffer a genuine loss without thereby acquiring an unlimited claim upon everyone else to make it good.
This is where sophisticated corruption lives.
The crudest thief falsifies the entries. The cunning predator manipulates the relationship between the books.
He converts legal permission into moral innocence.
She converts genuine victimhood into an unlimited entitlement.
An institution converts procedural regularity into substantive legitimacy.
A government converts a desirable collective objective into authority it was never granted.
The books can each appear internally correct while the consolidated accounts are fraudulent.
The crucial audit problem is therefore not merely whether the entries are correct, but whether the transformations between ledgers are legitimate. An immutable ledger can preserve a fraudulent transformation perfectly.
Corruption can therefore appear first as an illegitimate conversion between ledgers, long before anyone recognises it as a wrong, let alone a crime.
Perhaps that is how a criminal-industrial complex grows without most of its participants ever experiencing themselves as criminals. Each person maintains one ledger while assuming that somebody else is reconciling the whole.
Nobody is.
The final constraint is spiritual
And this is why even the most spectacular societal cleanup cannot finally solve the problem. We can replace institutions, prosecute criminals, expose fraud, decentralise authority and make transactions transparent. We can place records on immutable ledgers and deploy artificial intelligence to inspect them continuously.
But technology cannot finally decide what ought to be entered in the books, or which transformations between them are just.
A civilisation must cultivate people willing to record an entry against themselves; willing to recognise an obligation they could technically evade; willing not to exercise a power merely because they possess it; willing to distinguish restitution from revenge; willing to accept an adverse outcome when the process was fair; and willing to reopen the books when evidence shows that it was not.
Technology can make discrepancies visible. Law can specify permissible transformations. Institutions can reconcile competing claims. AI can audit the resulting machinery.
None can supply the desire to be just.
So perhaps the real task after the takedown is not constructing a world in which corruption becomes impossible. Such a world cannot exist while human beings remain human.
It is constructing a world in which the covenantal, contractual and cost-based ledgers cannot easily be forced apart—and cultivating people who do not want to force them apart in the first place.
The first is an engineering problem.
The second is a spiritual one.
We will need to solve both.
The state needs auditing.
The auditor needs auditing.
The rules need auditing.
The mappings between the ledgers need auditing.
And finally, the person holding the audit tool must be willing to audit himself.