The latest July data from the Road Safety Foundation (RSF) brought no surprises, only a familiar but tragic reality.
The organization counted 416 people killed and 629 injured in 458 crashes across Bangladesh. Coming right after June's toll of 438, the country lost 854 lives in just two months.
Yet, these shocking figures caused barely a ripple in the national conversation.
Somewhere inside those two months, at a United Nations high-level meeting on road safety in July, Bangladesh restated its commitments on halving road deaths and injuries by 2030 and enacting a comprehensive road safety law by December 2027.
The pledge was made abroad while the July toll was accumulating at home.
Around the time the March Eid-rush was filling our highways, a smaller and far less dramatic vehicle arrived in Singapore.
It is a 16-seater electric bus, studded with cameras and sensors that give it a complete 360-degree view of its surroundings. It has no steering wheel, no pedals, and no human driver who can lose patience, experience fatigue, or take a reckless move to overtake a rival on a blind curve.
Land Transport Authority (LTA) received the first of six autonomous buses that will enter public service as part of a three-year pilot in mid-2026. Before the bus touches a public road, it will be put through rigorous closed-circuit testing, a stark contrast to the untested, unfit vehicles routinely deployed on Bangladeshi highways.
The price of a deadly status quo
The contrast with our own roads is not subtle. According to RSF, at least 274 people were killed and over 1,500 injured in 342 accidents during the 10-day Eid-ul-Fitr travel window from March 16 to 26.
The deadliest single incident occurred when a passenger bus lost control and plunged into the Padma River from a pontoon at the Daulatdia ferry ghat, claiming 26 lives.
A mail train rammed into a passenger bus after sleeping gatekeepers failed to lower the security barriers, killing 12 more at a level crossing in Comilla.
By the end of March, the toll had climbed to 532 deaths and 2,221 injuries across 576 accidents during that month alone.
The fatalities continued unabated into April 2026, which recorded 404 deaths and 709 injuries from 463 crashes. This grim period seamlessly transitioned into the next holiday season.
During the 13-day Eid-ul-Adha travel period from May 21 to June 2, 2026, the country witnessed another wave of carnage. RSF reported 281 fatalities and 837 injuries in 292 road accidents nationwide during this rush, with motorcycles proving to be the deadliest mode of transport, accounting for 124 of those deaths.
Among the most devastating of these tragedies was an incident on May 25 in Tangail's Kalihati Upazila, where an overloaded cargo truck illegally carrying homebound passengers under a tarpaulin cover overturned into a roadside ditch, crushing 15 labourers to death.
These occurrences were far from unusual; they were the predictable outcomes of a system that has failed to protect people year after year, Eid after Eid.
Then the festivals ended, and the numbers did not move. That is the part worth sitting with. June and July carried no holiday rush and no seasonal excuse and still produced 854 deaths between them.
In July, motorcycle crashes killed 138 people, just over a third of the total. Pedestrians accounted for another 106. National highways, the stretches built for speed and supposedly the best engineered, hosted 169 of the month’s crashes. Loss of vehicle control was again the single largest cause.
A separate count published in late July, days before the eighth anniversary of the 2018 student road safety movement, found that 7,364 students had been killed on the roads between 2019 and June this year, more than 15% of all road deaths in that period.
Of those, 391 died in the first six months of 2026. The children who marched for safer roads in 2018 have been replaced on the casualty lists by children who were too young to march.
The overall picture is grim, with the RSF’s 2025 Eid datahighlighting the specific catalysts of the crisis. Loss of control caused 44% of accidents, and head-on collisions made up 26%.
Motorcycles, which quadrupled from 1.5 million to over six million in a decade, accounted for more than 42% of all Eid fatalities.
Ultimately, behind almost every statistic lies an avoidable human error: An exhausted driver racing against the clock on a blind curve, or a desperate motorcyclist riding without a license or a helmet.
What is the real count?
There is a fundamental failure sitting underneath all of this. Bangladesh, it seems, cannot accurately track its own road fatalities.
For 2025, the Bangladesh Road Transport Authority (BRTA) recorded 5,490 deaths and called it the highest annual figure on record.
RSF, working from media reports, counted more than 7,300.
The Bangladesh Jatri Kalyan Samity put it at 9,111.
The World Health Organization, using modelled estimates, once placed the toll for a single earlier year at close to 32,000 while local counts for that same year sat below 8,000.
These are not variations at the margin. Between the lowest and the highest domestic count sits a gap larger than the entire annual road death toll of many if not most countries.
An agency cannot target what it cannot measure. Blackspot identification, budget allocation, and any honest test of whether a policy worked all rest on a number the state does not reliably possess.
Singapore will monitor six autonomous buses in real time and judge them against a defined standard. We cannot say within several thousand how many of our own citizens died last year, and no institution carries an obligation to reconcile the difference.
The inspection machinery is similarly thin. Of five vehicle inspection centers in Dhaka, officials have acknowledged that only one is operational. Seven new centres are planned across the divisions under a World Bank-funded program running to 2030, which is welcome and is also an admission of how little exists today.
When a directive goes out that unfit vehicles must come off the roads, the arithmetic defeats it before corruption gets a chance to. The physical capacity to inspect the fleet is simply not there.
Beyond the breaking point
This is where Singapore's autonomous public bus becomes more than a curiosity. By no means, it’s being proposed as an immediate solution for Bangladesh. Our chaotic road conditions, fractured regulatory frameworks, and complex traffic environment wouldn’t allow a similar implementation anytime soon.
However, the guiding philosophy of the machine is what has been missing from our road safety framework all along. The autonomous bus operates on a logic of absolute rules. It cannot be pressured by an owner to squeeze in one more trip, nor can it be bribed to ignore a red light. It does not have patrons in the administration to call when pulled over. It follows the rules because it has no mechanism for exception.
Calling these fatal road disasters "accidents" shifts responsibility to fate. Experts accurately label them as “negligent killings,” the direct consequence of a transport culture defined by impunity.
At least 6.26 lakh unfit vehicles are currently operating across the country with the tacit compliance of regulators. Furthermore, the vast majority of commercial drivers work without formal employment contracts, fixed operational hours, or fair wages.
This institutionalized exploitation creates a state of chronic fatigue and mental distress, conditions that RSF identifies as catalysts for the recklessness witnessed on our highways.
The transport syndicate is no shadow actor; it operates as a powerful institutional roadblock, ensuring that independent accident investigations are buried, identical fatal causes recur annually, and structural reform never takes hold.
Fundamentals over future-tech
The useful lessons from Singapore cost far less than a driverless fleet.
Start with the surveillance that already exists. The CCTV, speed-detection, and automated number plate recognition systems already installed along the Dhaka-Chittagong highway should be extended nationwide and wired to penalties that issue automatically, without an officer at the roadside deciding whether to write the ticket.
The government has spoken publicly of expanding AI-enabled speed management, so the intent is already on record; what is missing is coverage and automatic consequence.
Fatigue needs a legal instrument rather than an appeal to conscience. Digital tachographs, which log speed, distance and rest periods without needing a driver’s cooperation, should be mandatory on long-distance buses and trucks, with the financial penalty for exceeding legal hours falling on the company rather than on the man at the wheel.
Liability must move up the chain. An owner who knowingly puts a structurally compromised vehicle into service should face consequences that outweigh what the vehicle earns, and fitness certification should be digitized end to end so that a certificate records an inspection that really happened.
That, of course, requires inspection capacity to exist, which returns us to the seven centres and the years it will take to build them. Biometric licensing and mandatory safety training for the country’s motorcyclists, and for the battery-run three-wheelers now spreading across regional roads, belong in the same program.
Someone has to be in charge
Road safety responsibilities in Bangladesh are currently fractured across the BRTA, the home ministry, the Roads and Highways Department, and the railway and municipal authorities, which is why the Comilla inquiry had to implicate six separate offices before it could describe a single crash.
Singapore’s LTA manages everything from licensing and vehicle standards to infrastructure and enforcement under one roof, which is why it can attempt an autonomous pilot at all. Bangladesh needs a unified road safety authority with statutory power and a standing obligation to publish binding annual reviews.
None of this is unaffordable. The Road Transport and Highways Division has put the annual cost of crashes at around Tk 60,000 crore, with the BRTA estimating the burden at 1.08% of GDP for the last financial year.
International assessments suggest that an annual investment of roughly 0.2% of GDP could save on the order of 11,000 lives a year in Bangladesh. We are already paying several times that figure in wreckage.
Choosing discipline
The bus in that Singapore test facility represents a place where removing human error from transport is treated as an engineering problem worth the expense of solving properly.
Singapore made road safety a governance priority decades before any of this technology existed, and the autonomous pilot is simply where that long discipline has arrived.
Bangladesh has shown it can rally comparable conviction when it chooses to; the bridges and power plants of the past decade were not built by accident.
On the roads, though, the same families keep grieving the same preventable deaths while the same institutions produce the same reports and shelve the same recommendations.
The pledge made in July commits us to halving road deaths within four years. On the present calculation that means roughly 200 fewer people dying every month, beginning fairly soon.
It will not come from teaching buses to drive themselves but from counting our dead honestly, and from accepting that road fatalities are neither inevitable nor negotiable.
Dr Sabbir Ahmad, a tech executive with global experience in digital connectivity, sustainable infrastructure and energy, now shaping Bangladesh's semiconductor landscape as the CEO of Silicon Array Ltd. Email: [email protected].
