Bangladesh's ready made garment sector employs close to four million people and earns over 80% of the country's export revenue. For four decades, it has been the country's single biggest source of formal work outside the farm. That fact has held so steady for so long that we treat it as permanent. It is not.
The Centre for Policy Dialogue warned this month that up to 1.22 million RMG jobs could disappear by 2041 as automation spreads through the sector, with as much as 60% of current female employment at risk of being wiped out entirely.
The same report found something quieter and more worrying underneath that number. Manufacturing employment across the country has stayed frozen at 8.1 million workers even as production has climbed year after year. The factories are producing more. They are simply not hiring more people to do it.
This is not a distant scenario. Nearly 1.3 million jobs faced cuts throughout Bangladesh in 2024. 90% of that loss fell straight on women, according to official labour data cited in a recent industry report. The report's central complaint was not that automation is happening. It was that no policy exists to manage what happens to the people it displaces.
We are not the only country facing this. Vietnam's textile and garment workers face an even sharper automation risk, with the ILO estimating 86% of jobs in the sector at high risk from robotics and machine driven production.
Vietnam pays its garment workers roughly $300 a month against Bangladesh's $95, so it has less room to compete on labour cost and more reason to compete on output per worker.
Factories there report that automation now lets one worker produce four to five times what they produced before, and the country is pouring money into training programs so its workforce can run and maintain that equipment rather than be replaced by it.
That is the difference between agility and drift. Vietnam is not avoiding automation. It is choosing where automation goes and making sure its own people are the ones operating it.
Bangladesh has largely let automation arrive on its own terms, factory by factory, with no national plan attached to it.
There is genuine opportunity sitting inside this technology if we can capitalize upon it early. A study on AI adoption in Bangladesh's RMG sector has come to a noteworthy finding that real gains are available through better defect detection, demand forecasting, and production planning, the kind of AI applications that raise a factory's output without necessarily cutting its headcount.
The Daily Star has reported that AI powered systems in Bangladeshi factories are already predicting equipment failures before they cause costly downtime and catching stitching flaws faster than a human inspector could. Used this way, AI does not remove the worker from the equation. It removes the waste sitting around her.
But the same study found something that should worry us. Researchers identified a chronic gap between what factories are adopting and what regulation, training systems, and worker protections have prepared for.
There is no clear framework for accountability when an algorithm makes a production decision that costs someone their shift. There is no established path for a machine operator to become a machine technician. The technology is arriving at pace but we are not running them through a strategically agile phase.
Here, strategic agility should mean factories adopting AI for quality control and planning while simultaneously funding the retraining of the workers whose old tasks disappear.
It means vocational training centres building curricula around machine operation and AI assisted production today, not after the job losses have already happened.
It means BGMEA, the labour ministry, and training institutions sitting at the same table before the layoffs, not after them, so that a woman doing manual quality checks this year is doing machine assisted quality control next year instead of leaving the workforce altogether.
None of this is unfamiliar territory for Bangladesh. Our history tells us that we know how to move when the underlying economics forces our hand. AI is forcing our hand right now. The choice now is whether we plan for the workers standing next to it, or wait until the next report tells us how many of them are gone.
Nafis Ehsas Chowdhury works at M M Ispahani Limited.
