When the Algorithm Waits: The Six-Month Silence That Exposes Guyana’s AI Enforcement Illusion

Opinion | The 592 Guardian

There is a particular kind of deception that does not lie outright. It does not fabricate facts or invent events. It operates instead through selective emphasis — parading the gleaming face of progress while quietly shielding its contradictions from scrutiny. Guyana’s rollout of artificial intelligence in traffic enforcement has become a masterclass in exactly this kind of deception.
We are told that the system is sophisticated. We are told it is modern, efficient, and — most importantly — fair. And yet, 1,600 drivers identified by that very system as suspected violators waited six months to receive so much as a notification letter. Six months. In a world where artificial intelligence can scan a vehicle travelling at highway speed, cross-reference its plate against a national database, and generate an enforcement flag in fractions of a second, it apparently cannot send a letter in less than half a year.
That contradiction is not a minor administrative footnote. It is the story.

The Promise That Built the Narrative

To understand how deeply troubling this delay is, one must first appreciate the scale of what was promised — and what was sold to the Guyanese public.
AI-powered enforcement was not presented as a modest upgrade to existing traffic systems. It was framed as a transformational leap — a decisive break from the era of arbitrary roadside stops, inconsistent policing, and enforcement that depended too heavily on the discretion, and sometimes the appetite, of individual officers. The technology would be neutral. It would be tireless. It would see everything and treat everyone the same.

The language used in official communications about the system was careful but unmistakable in its ambition. Real-time detection. Automated flagging. Instant database integration. These were not the words of a department piloting a modest tool — they were the vocabulary of a government staking its modernization credentials on a technological promise.

That promise had real political currency. In a society where the belief that enforcement is selective — that who you know determines whether the law applies to you — remains deeply entrenched, the idea of a machine-driven system carried genuine appeal. A camera does not accept a handshake. An algorithm does not respond to a phone call. If the technology is what it claims to be, then the rules truly do apply to everyone.
That was the bargain offered to the public. The six-month silence is proof that the bargain was not honoured.

What Artificial Intelligence Actually Does — and Doesn’t — Explain

It is worth being precise here, because the government’s defenders will reach for a familiar rebuttal: that even with AI detection, downstream enforcement processes involve human steps that take time. Letters must be drafted, addresses verified, decisions reviewed. Technology, they will argue, does not eliminate the need for administrative procedure.

This argument is not entirely without merit. It is, however, entirely insufficient.
Artificial intelligence, at its core, is a tool for compressing time. It eliminates the bottlenecks that arise when human judgment must be applied to each individual case. A system that can flag 1,600 violators does not require 1,600 separate human decisions to generate 1,600 notification letters. That is, in fact, the entire point. A functioning AI enforcement system would have the capacity to automate the notification pipeline with the same efficiency it applies to detection. If such automation was not built in — if the detection engine was wired to a manual, paper-shuffling bureaucracy on the back end — then the system as deployed is not what it was described to be. It is a camera with a filing cabinet.

More pointedly: if the technology is capable of triggering enforcement responses on the spot — if, as has been suggested, it can interface in real time with officers in the field — then the claim that notifications to flagged drivers required six months of processing collapses under the weight of its own implausibility. You cannot tell the public that your system can alert a patrol officer to a flagged vehicle in motion, and then explain that generating a mailed notification to the same vehicle’s registered owner requires half a year.
One of these claims must give way. And the public is entitled to know which one.

The Silence That Speaks

Six months is not a delay. It is a decision.
In public administration, especially in matters of law enforcement, timelines are never purely accidental. Resources are allocated. Priorities are set. When an action that should be routine — notifying individuals that they have been flagged for a potential offence — takes six months to execute, it is because something, or someone, chose not to expedite it.
The question that follows from that observation is the one the government has been conspicuously reluctant to answer: what happened during those six months?

Who held the list of 1,600 flagged drivers? Who had access to it? Was it reviewed by anyone outside the technical enforcement team? Were any names removed before notifications were issued? Were any names added? Were the 1,600 who eventually received letters the same 1,600 originally flagged by the system — or had the list been quietly adjusted along the way?

These are not hypothetical concerns born of cynicism. They are the natural and reasonable questions that arise any time sensitive enforcement data is held in bureaucratic limbo without accountability or transparency. The longer the silence, the more weight those questions carry.
And they are carrying considerable weight.

The Whispers That Grow into Walls

Across Guyanese civil society, in the conversations that happen between citizens rather than in official briefings, a particular suspicion has taken root. It is not yet a verified allegation. It has not been confirmed by any whistleblower or official source. But it has spread with the stubborn persistence of ideas that feel intuitively true to the people who hold them.
The whisper is this: the delay was not administrative. The list required “scrubbing.” Certain names — names attached to individuals with the right connections, the right relationships, the right proximity to power — were quietly removed before the letters went out.

Again: this has not been proven. It may not be true. But the extraordinary danger of this moment is that, true or false, the suspicion is entirely plausible given what the public has been shown of how enforcement works in Guyana. And a government that has spent years promising a new, transparent, technology-driven era of equal treatment has done nothing — nothing — to proactively foreclose that suspicion.

No audit of the flagged list has been published. No independent verification of the notification process has been announced. No explanation for the delay that goes beyond vague administrative reference has been offered. In the vacuum of credible official explanation, the whisper does not merely survive — it thrives.
This is the corrosive power of perceived selective enforcement. It does not require proof to do its damage. It only requires the absence of transparency. And on that front, the government has been remarkably, perhaps recklessly, generous.

Artificial Intelligence Cannot Survive Artificial Fairness

There is a term that deserves wider circulation in this debate: artificial fairness.

---

Artificial fairness is what you get when you deploy a genuinely neutral technology — a system that, left to its own outputs, would apply the same standard to every vehicle that passes before its sensors — and then introduce human discretion back into the process after the fact. The detection may be real. The flagging may be real. But if what comes out the other end of the pipeline has been filtered through human judgment that is susceptible to influence, then the neutrality of the machine is merely cosmetic.

---

This is not a hypothetical risk. It is the specific vulnerability that the six-month delay has exposed. AI systems produce outputs. Those outputs then enter a human-administered pipeline. If that pipeline operates without transparency, without independent oversight, and without verifiable timelines, then the AI’s integrity guarantees nothing. You have simply moved the point of potential manipulation further downstream, where it is harder to see and easier to deny.

The administration has built its modernization narrative on the premise that AI enforcement represents a structural departure from the culture of discretion. But discretion does not disappear when a camera is installed. It migrates — to the person who controls the database, to the official who reviews the flagged list, to the bureaucrat who decides which letters go out and when. If those human nodes in the enforcement chain are not bound by the same transparency and accountability standards applied to the technology itself, then the system as a whole is no more trustworthy than its weakest human link.

Right now, that link is invisible, and it has been invisible for six months.

The Standard That Must Be Applied

What would credible AI-powered enforcement actually look like? It would look like this:
Detection events would be logged in a tamper-evident, time-stamped system that is subject to independent audit. The interval between a flag being raised and a notification being issued would be defined in advance, published publicly, and enforced consistently. Any deviation from that interval would require a documented justification, accessible to oversight bodies. The composition of any flagged list — who is on it, when they were added, and whether any names were ever removed — would be auditable by a body that is genuinely independent of the enforcement apparatus.

None of this is technically complicated. All of it is politically demanding. It requires a government willing to be held accountable not just to its own stated standards, but to external verification of whether those standards are being met.
The question before the Guyanese public is not whether AI enforcement is, in theory, a good idea. It may well be. The question is whether the specific deployment currently underway can be trusted — and whether the administration responsible for it is willing to do what trust requires.

At present, the evidence suggests the answer is no.

Modernization Is Not a Marketing Exercise

Guyana is at a genuine inflection point. The country’s economic transformation over the last several years has created real capacity for institutional modernization. Resources exist that did not exist before. The appetite for a more functional, more equitable state apparatus — particularly among younger Guyanese — is real and should not be squandered.

But modernization is not achieved by acquiring technology. It is achieved by building institutions capable of deploying technology in ways that genuinely serve the public interest. An AI camera on a highway is hardware. The culture that governs what happens to its outputs is the institution. And it is the institution — not the hardware — that determines whether the system produces justice or merely produces the appearance of it.

A country that installs cutting-edge detection technology and then subjects its outputs to an opaque, unaccountable, six-month human filtering process has not modernized. It has digitized its old habits. It has made them faster, more scalable, and — if the technology’s reputation for neutrality is successfully leveraged — considerably harder to challenge.

That is not progress. That is the old order in new clothes.

The Questions That Cannot Be Deferred

The administration owes the public answers. Not reassurances — answers. Specific, verifiable, documented answers to questions that are neither unreasonable nor hostile. They are the questions that any functioning democracy asks of its enforcement apparatus:

Why were the 1,600 flagged drivers not notified within days of being identified? What is the documented justification for the six-month interval?

Who had access to the flagged list between the time it was generated and the time notifications were issued? Was that access logged?

Were any names removed from the original flagged list before notifications were sent? If so, on what authority, under what criteria, and with what documentation?

What independent oversight body has visibility into the enforcement pipeline, and what are its powers to audit, challenge, or publicise its findings?

Until those questions are answered — publicly, specifically, and with supporting documentation — the rollout of AI-powered traffic enforcement cannot be accepted as the transparent, technology-driven modernization it has been presented as.

It remains, for now, a performance. An expensive, technically impressive, and politically convenient performance — but a performance nonetheless.

What Guyana Deserves

The people of Guyana deserve enforcement that is credible. Not enforcement that is claimed to be credible. Not enforcement that is credible in its technology while opaque in its administration. Credible in the full and demanding sense: where the rules apply equally, timelines are consistent and published, the data is protected from interference, and the institutions responsible for it are genuinely accountable to the public they serve.

That is not a utopian standard. It is the baseline expectation of a functioning rule of law. Other countries meet it. Guyana can meet it too.

But it will not be met by installing cameras. It will be met by the hard, unglamorous, politically costly work of building institutions that cannot be quietly negotiated with — where the algorithm’s output is as binding on the well-connected as it is on everyone else, and where the word “enforcement” does not carry an asterisk.

Anything less is not modernization. It is not efficiency. It is not justice.
It is digitized inequality — and it insults the intelligence of every Guyanese citizen who was told the machine would be different.

𝙏𝙝𝙚 592 𝙂𝙪𝙖𝙧𝙙𝙞𝙖𝙣 𝙞𝙨 𝙖𝙣 𝙞𝙣𝙙𝙚𝙥𝙚𝙣𝙙𝙚𝙣𝙩 𝙂𝙪𝙮𝙖𝙣𝙚𝙨𝙚 𝙘𝙤𝙢𝙢𝙚𝙣𝙩𝙖𝙧𝙮 𝙖𝙣𝙙 𝙤𝙥𝙞𝙣𝙞𝙤𝙣 𝙤𝙪𝙩𝙡𝙚𝙩 𝙘𝙤𝙫𝙚𝙧𝙞𝙣𝙜 𝙘𝙞𝙫𝙞𝙘, 𝙥𝙤𝙡𝙞𝙩𝙞𝙘𝙖𝙡, 𝙖𝙣𝙙 𝙧𝙚𝙜𝙞𝙤𝙣𝙖𝙡 𝙖𝙛𝙛𝙖𝙞𝙧𝙨.