[🇧🇩] Artificial Intelligence-----It's challenges and Prospects in Bangladesh

[Saif]

AI opens new revenue stream for telcos

For decades, telecom operators built their businesses around one basic product -- connectivity. They first sold voice calls, then gradually turned mobile data into a much larger source of revenue as smartphones and mobile internet transformed how people communicate, work and consume digital...

www.thedailystar.net

AI opens new revenue stream for telcos

Mahmudul Hasan

Telcos shifted from voice to data, and now AI services. Photo: Star/file

For decades, telecom operators built their businesses around one basic product -- connectivity. They first sold voice calls, then gradually turned mobile data into a much larger source of revenue as smartphones and mobile internet transformed how people communicate, work and consume digital services.

Worldwide, telecom operators are beginning to sell something different: access to artificial intelligence. Bangladesh’s mobile operators are moving in the same direction.

Robi introduced Robi AI Space on September 9, offering access to ChatGPT, Gemini, Claude and DeepSeek.

A day later, Grameenphone launched OneAI, bringing multiple leading AI models under a single subscription through the MyGP app.

Banglalink, meanwhile, has been offering AI services through its RYZE platform, including access to multiple models and AI token packages.

The moves signal a shift in the operators’ role -- from simply providing the connectivity through which digital services are consumed to becoming distribution, subscription and payment platforms for AI services.

Globally, the development is part of a broader shift in the telecom industry from selling network traffic to monetising AI consumption.

China’s three major telecom operators -- China Mobile, China Unicom and China Telecom -- launched commercial AI-token packages in May, marking one of the clearest examples of mobile operators treating AI usage as a product similar to traditional voice minutes or data allowances.

So what exactly are these AI tokens?

Tokens are the basic units used by large language models to process information. When a user sends a prompt to an AI service, the input is broken down into tokens, while the model’s response also consumes tokens. The amount of processing required can therefore be measured in tokens, allowing AI providers to charge according to usage.

In simple terms, a mobile operator sells gigabytes to pay for internet consumption, while an AI platform can sell tokens to pay for AI computation.

China’s operators have begun packaging those tokens much like mobile data. China Telecom, for example, launched consumer plans starting at 9.9 yuan ($1.40) a month for 10 million tokens, with higher-priced packages offering larger allowances. It also introduced business and developer packages, with token allowances reaching hundreds of millions.

China Unicom also launched token packages, including plans ranging from six million to 18 million tokens, while China Mobile has offered token-based services through partnerships and local branches.

That gives operators an opportunity to move beyond their traditional role as connectivity providers.

The model is now reaching Bangladesh.

Tanveer Mohammad, chief corporate affairs officer at Grameenphone, said that when voice service first arrived, it brought a significant transformation to people’s lifestyles and engagement at that time. Then came the era of data. By using data, people gained immense access to global knowledge.

“Now, the entire world is entering the era of AI. Telcos cannot stay outside of this global AI revolution. We must leverage AI to increase our operational efficiency, expand our capabilities, and broaden our reach.”

Grameenphone’s service includes ChatGPT, Claude, Grok, DeepSeek and Gemini, among other AI models. Customers can use the service for learning and research, content writing, coding and software development, image generation, document analysis and summarisation.

At an event at GPHouse marking Grameenphone’s AI&I Day, Chief Guest Rehan Asif Asad, adviser to the prime minister for telecom and ICT, officially announced the launch.

Rehan Asad said, “I am encouraged to see Grameenphone taking the lead in making AI more accessible to people through initiatives such as OneAI. For students in particular, this creates an affordable way to access AI tools and use them to support learning, research and skill development.”

Grameenphone CEO Yasir Azman said, “We are putting the possibilities of AI directly into the hands of millions of our customers, particularly students, making it simpler and more accessible for learning, work, creativity and problem-solving.”

Through AI Space, Robi positioned the service as an all-in-one AI hub inside the My Robi app.

Through a partnership with Every AI, Robi AI Space offers access to ChatGPT, Gemini, Claude and DeepSeek through a single premium subscription or selected AI bundle packages.

Robi Managing Director and CEO Ziad Shatara said, “With credit card penetration posing a structural hurdle, our market has long faced restricted access to global digital tools. By introducing this all-in-one AI platform through My Robi, we are bridging that divide.”

Banglalink has also entered the emerging AI-services market through RYZE, its AI-powered digital platform.​

 

[Saif]

What will remain ours in the age of AI?

In the past, people would view a glacier chunk falling into a lake and causing a river to overflow as a cosmic event or divine act.

www.thedailystar.net

What will remain ours in the age of AI?

Shamsad Mortuza

VISUAL: SALMAN SAKIB SHAHRYAR

In the past, people would view a glacier chunk falling into a lake and causing a river to overflow as a cosmic event or divine act. Now that we have placed an eye in the sky, the satellite images can help us understand how the unprecedented flash flood in the Nepal-Tibet border area happened. And to aid human reason, AI is employed to piece together the exact way the mountain and the river interlocked in a cataclysmic event.

The event is a reminder how small and insignificant humans are in the grand scheme of things. Many images online highlight humanity’s extremes: people risking their lives to save others, while some others scavenging through debris to loot from the dead. One image of a two-storey building surviving the onrush of waters made many of us think of miracles. Then came the scientific explanations, where structural engineers showed the extra piling used to anchor the house. The image made me reflect on the very connection between deep and surface structures.

While presenting a plenary talk at the recently held AI conclave at BRAC University, I used this solitary house image as a metaphor to argue for the need for a strong foundation in a deluge of information. As a student of literature and culture, I have a holistic approach to the onrush of information that has flooded our screens. In academia, we are unsure how to deal with the exponential growth of technology. There is no denying the fact that research today is much more computational, data-driven, industry-led and interdisciplinary; knowledge production relies heavily on technology. AI is here to stay, but how to approach it is another matter.

An MIT media lab publication titled “Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task” finds that brain-only writers have the strongest, followed by search users with intermediate, and Large Language Model (LLM) users with the weakest brain connectivity. Yale University has responded by developing an AI guidance for students, encouraging them to think critically about the technology.

As policymakers have stated sporadically, Bangladesh is working on an AI policy. AI is now being considered a life skill. Algorithmic ads targeting groups of people becoming experts on digital tools in under a month are plentiful. But can one truly become an expert in anything without a solid grounding in that subject? That is precisely the question I floated in the AI conclave. I can write a prompt to ask the LLM to produce a poem for me. But can I claim it as my own? Do I own the language that exists in its repository? This is where the issue of ethics and integrity comes in.

A culture of suspicion has arisen where we begin to comprehend the success of others as a compromised integrity. We think the singers have used auto-tuners, academics have used AI, and business houses have used manufactured data for their sales pitches. This is a culture where some assume a moral high ground to preach honesty and integrity.

It is like students turning in an assignment not realising that the teacher has added a hidden catchphrase to nab a misconduct. On Teachers’ Day, ChatGPT released a short video to celebrate teachers by saying, “We know.” The ad was supposed to tell the students not to think that they had outwitted their teachers by using LLM tools. Teachers have the foundational knowledge to know that their students are cutting corners. The temptation is too powerful. We are all beginning to outsource our tasks: students finishing assignments in 15 minutes and teachers preparing their lesson plans or class materials in equally short time. Information is processed at an unprecedented scale and speed. Our overreliance on devices has made it increasingly difficult to think on our own. We are rehearsing knowledge that is already in the repository, then producing it with some personal garnishing.

A representative from the leading plagiarism checker Turnitin told the conclave how the number of submissions to Organization Science had increased by 42 percent ever since ChatGPT was launched in 2022. Even the prestigious publishing outlet like Springer Nature retracted 2,923 papers in 2024. A Ukrainian paper mill network was found responsible for 1,517 fraudulent papers. On September 8, TBS Bangla ran an investigative report and talked to a Bangladeshi paper mill selling Q1 journal articles for TK 20,000-70,000.

So how do we pursue knowledge and retain trust and integrity in this age of AI? More importantly, how do we retain our human voice and agency at a time when everything can be outsourced? I return to the image of the house that survived Nepal’s flood. We need to build a foundation before we encounter the flood of information presented by AI. And that foundation must be built in the right environment.

We need to be AI-resilient, but thinking that we can be AI-proof will be foolish. Instead, we can adopt a product-process-presence model. If we consider an assignment to be a product, for instance, we need to check it for accuracy, disciplinary standards, and the evidence that it provides. Then there should be declarative options where the producer of the work must reveal how the product was made. Some journals now require the proposals, drafts, sources, prompt log and revisions made. The most important phase is the presence phase: where a learner must be able to defend the product, maybe through a viva or a brief presentation.

For such an ecosystem to grow, we need to apply the same standards to both teachers and students. Teachers must make the AI rules clear. Students need to know whether AI is completely prohibited or partially allowed. Whatever the case is, the feedback process must retain human agencies.

At an institutional level, policymakers must not create unrealistic pressures on their faculty members in the name of rankings and reputations to publish beyond their capabilities. The pressure chain prompts faculty members to salami-slice their works, acting like the proverbial fox repeatedly showing the same crocodile kid before consumption.

There is no glory in catching plagiarism after it has happened. The cop-and-robber game that the media likes to dress up as news is just the tip of the iceberg. The greater problem is our knowledge that the glacier is heating up. Whatever we know as our own will have a great fall. And if we all follow the fancied trend, we risk being washed away. The only way we can withstand the change is by making us ready for the next flood. AI can help us build faster, but only we can decide what is worth building and where we want to build next.

Dr Shamsad Mortuza is vice-chancellor at the University of Liberal Arts Bangladesh (ULAB).​

 

[Saif]

https://www.newagebd.net/post/science-and-technology/313911/rapid-development-of-ai-crisis-or-opportunity-for-bangladeshi-youth

Rapid development of AI: Crisis or opportunity for Bangladeshi Youth?
Porimol Palma 15 September, 2026, 11:31

Artificial intelligence is rapidly changing the way people work, from writing and translation to coding, data analysis, graphic design and software development.

For Bangladesh, however, the transformation is arriving at a difficult time: millions of young people are already struggling to find work.

The Bangladesh Bureau of Statistics Labour Force Survey put the country's overall unemployment rate at 3.66 per cent in 2024. But unemployment among people aged 15–29 was 8.07 per cent.

The rapid spread of AI therefore raises a critical question for Bangladesh – if whether it will create new opportunities or make an already difficult job market even harder to enter.

For Achia Nila, founder of software development company Women in Digital, AI is both a challenge and a major opportunity.

AI, she said, is no longer relevant only to IT professionals. People in almost every sector can use it to improve productivity.

She herself has seen the change in her business.

‘I used to employ ten developers before the advent of AI. Now, I can do much of my primary work, including project analysis, myself as I have an IT background,’ she said.

‘So, I might have cut some employees, but the efficiency and speed of my work has have been enhanced.’

That experience captures one of the central dilemmas posed by AI. The technology can make an individual worker significantly more productive, but the same productivity gain can mean that companies need fewer people to perform certain tasks.

For Bangladesh's large pool of freelancers, particularly those providing services such as data entry, graphic design and basic software development to overseas clients, the implications could be significant.

‘There is a common perception that the Bangladeshi freelancers engaged in providing IT products and services to the developed world will be hardest hit by the advent of AI,’ Nila said.

‘This is a partial truth. If you update yourself with the required expertise, you would be able to work more efficiently and earn even better.’

The bigger danger is for workers who do not update their skills.

Continuous learning and adaptation is are increasingly important as technology changes both the nature of work and what employers expect from workers.

‘AI cannot work without humans behind them. It needs to prompt and it is the humans who can be prompters,’ Nila said.

Even AI-generated content, she added, requires human supervision and verification.

People are already seeing large numbers of AI-generated images and videos on social media, but the novelty of such content is beginning to wear off.

‘People want unique, human-touched contents content. Even if you take AI assistance, it has to have unique features, and that’s not possible without humans working behind them,’ she said.

Entry-level jobs under pressure

Professor Rifat Shahriyar of the Department of Computer Science and Engineering at BUET has observed a more immediate impact on the country's IT job market.

‘The demand for the fresher CSE graduates has come down because of the industries’ more use increased use of AI,’ he said.

Starting salaries for newly recruited graduates have also declined, he said.

Companies are increasingly using AI to enable mid-level engineers to perform software development and other tasks that previously required larger numbers of entry-level workers.

‘In the developed world too, the expert-level jobs on the subject are increasing,’ Shahriyar said.

That has implications beyond the technology industry because entry-level employment traditionally provides young graduates with their first opportunity to gain professional experience.

If those first rungs of the career ladder disappear or become fewer, young people may find it harder to enter the professions in the first place.

The same pressure could eventually affect freelancers working in routine areas such as graphic design and data entry.

But workers who learn to use AI effectively could find new opportunities, Shahriyar said.

A warning for Bangladesh's education system

A recent study of university students in Bangladesh suggests that young people are aware of the transformation but are not uniformly prepared for it.

The study, published on September 10 in Discover Global Society, surveyed 450 students at Dhaka University, Rajshahi University and Pabna University of Science and Technology.

The students generally believed that AI could create new employment opportunities, although they also expressed concerns about job losses and skill gaps.

The researchers called for greater integration of AI-related education into university curricula and closer alignment between what students learn and what industries require.

That challenge comes when, as per BBS Labour Force Survey 2024, there are according to the BBS Labour Force Survey 2024, there were about 2.62 million unemployed people in Bangladesh. Around 885,000 of them were university graduates, while unemployment among people with tertiary education was 13.54 per cent.

Bangladesh as AI consumer

Professor Shahriyar said the problem in Bangladesh is that it is basically the consumers a consumer of AI tools, not producers a producer. Also, the country does not have its infrastructure ready.

‘We are now only buying the applications or using free ones. We will be spending huge foreign currency amounts of foreign currency if we continue to do this.’

The country needs stronger domestic capacity to develop AI applications and supporting infrastructure, including data centres, he said.

‘We have talents talent and if we can develop our own applications, our government and industries can use them at competitive cost.’

Developing domestic capacity would also help Bangladesh address data protection and cybersecurity concerns, he added.

The issue is becoming increasingly important as AI spreads beyond technology companies into government, education, finance, healthcare, manufacturing and other sectors, the academic added.

Training the whole workforce

Both experts argue that AI training cannot be restricted to computer science students or software professionals.

‘The most important thing now for our government is to expand AI training to anyone eligible and have laws so that we cannot abuse AI and protect the data,’ Shahriyar said.

‘AI is not about the IT professionals only but for all sectors. It affects all and benefits all. What is needed is the right policy and regulations.’

The technology could also become an important learning tool.

People can use AI to learn foreign languages, mathematics and other skills, provided they know how to use the technology properly and verify the information it provides.

Nila similarly believes that Bangladesh needs to move much faster.

‘You can have your AI agents that can work for you if you train them properly. These are also paid AI services. You don’t have much of them many of them being used in Bangladesh,’ she said.

That points to a potentially important divide in the future labour market: not simply between people who have jobs and those who do not, but between workers who know how to use AI and those who do not.

The opportunity behind the disruption

Globally, AI companies are developing increasingly powerful systems at extraordinary speed.

Some technology leaders and researchers have warned that the development of advanced AI needs stronger safeguards because of the potential economic, cybersecurity and social risks.

But the technology also has the potential to dramatically increase productivity.

For Bangladesh, the challenge is therefore not whether AI can be stopped. It cannot.

The real challenge is whether the country can prepare its young population quickly enough.

‘We need strong regulations for data collection and incentives for people and industries to learn the required skills and develop local applications and infrastructure,’ said Professor Rifat Shahriyar.​

 

[Saif]

Back to '80s courtesy of AI!

Humans always have an inner urge to go back to the past, even though it is known that the clock cannot be reversed in the real world. This is reflected in the latest trend of generating photos from the 1980s using artificial intelligence (AI). Suddenly, many people have joined the trend to recreate

thefinancialexpress.com.bd

Back to '80s courtesy of AI!

Humans always have an inner urge to go back to the past, even though it is known that the clock cannot be reversed in the real world. This is reflected in the latest trend of generating photos from the 1980s using artificial intelligence (AI). Suddenly, many people have joined the trend to recreate their photos in '80s style. They are also sharing these photos alongside current ones on social media. The trend has also sparked debate about the extent to which AI recreates photos to resemble the '80s.

Several AI image generators are now in use by millions of people across the world to recreate their photos. These image generators recreate the photos mainly in two ways. One is using default settings, where people only need to upload their current photos to AI sites to get photos with '80s backgrounds, colours, and styles. Another is using text prompts to recreate uploaded photos so the AI can generate more unique photos for individuals. The more precise the prompt is, the better the photos. For instance, instead of asking to convert a photo into an authentic 1980s photo, it would be better to split the request into specific elements like 'the objects in the scene, clothing, photographic technology, cultural setting and the circumstances surrounding the photograph.' Several media outlets are now publishing tips on how to give precise prompts to AI for creating the desired photos.

Though Bangladeshis are also asking AI to generate photos of the '80s, in most cases what they are getting is photos having a flavour of the decade in the Western world, America to be precise. This is because AI's backend data are mostly loaded with photos, videos and descriptive text from the Western world. Some also include photos of social and fashion trends in India during the decade, especially Bollywood styles. Thus, Bangladesh in the '80s is largely absent from these AI-generated photos.

The 1980s were a time when Bangladesh confronted Ershad's autocratic regime to restore democracy. Old photos from those days show walls of public universities and colleges covered with slogans like: samorik ain tule nao, gonotantra phirye dao (withdraw the martial law, reinstate democracy). It was the decade when the export-oriented ready-made garments (RMG) industry began its journey to change the country's economic pattern. The advertisement of Pearson's: aaj ebong agamaikaler poshak (clothing for today and tomorrow).

In this decade, private entrepreneurship got recognition for the first time, after scrapping the state-driven socialist approach of the '70s. It was also the decade when unskilled and semi-skilled workers from Bangladesh started to go to the Middle East. Dubai jamu taka dao (I will go to Dubai, give me the money) was a line from a popular TV drama that spread by word of mouth across the country, even when there was no internet. In today's social media terms, the thing was viral. Rotary dial telephone set was a matter of social status at that time as only a few households had the telephone connections.

Again, the '80s was the decade of flourishing two-weekly magazines: Bichitra, edited by Shahadat Chowdhury, and Jai Jai Deen, edited by Shafiq Rehman. The latter faced a ban under the Ershad regime. In the decade, paperback thriller fictions and translations from Sheba Prokashoni, founded by Qazi Anwar Hussain, reached their peak. Almost every railway station bookstall sold Sheba's books. In the '80s, listening to the BBC Bangla service on the radio in the evening and morning was quite popular, especially when political confrontation occurred. BTV's weekly dramas and soap operas attracted the audiences to the screens. Watching movies at cinema halls was also a popular form of entertainment and socialisation. Razzak-Bobita, Alamgir-Shabana, Sohel Rana-Suchorita dominated the big screen. Young couples exchanged their photo view cards as a way to express their love to each other. Songs by Kurshid Alam, Abdul Jabber, Sabian Yasmin, Runa Laila, and Andrew Kishore were so popular that many recorded them on cassette reels. The craze for football, especially the rivalry between Abahoni and Mohammedan, was one of the most iconic symbols of the '80s. Salahuddin, Nannu, Salam, Chunnu, Badal, Aslam, and Kaiser were the great players whose photo posters lined the walls of many houses.

Thus, trying to go back to the '80s in Bangladesh using AI is nothing but a distorted effort. An AI-generated image of the decade can only provide a scattered interpretation, full of inaccuracies.​