Unregulated AI Boom Invites Disaster for Bangladesh

Recently, two telecommunication companies in Bangladesh launched services that allow people to easily access multiple Artificial Intelligence (AI) platforms, including ChatGPT, Claude, DeepSeek, Grok, and Gemini. Even the government is praising the initiatives, stating that “the role of AI cannot be overlooked as Bangladesh works towards its ambition of building a trillion-dollar economy by 2034.”

Adopting modern technology and automation is essential for the nation’s development; however, do the authorities fully grasp the impact of simplifying access to AI tools for citizens? There is no doubt that AI has enormous capabilities and value, but realising its benefits requires confronting and managing its pitfalls.

In Bangladesh, nearly half of children aged 7-14 still fail to read and understand a simple story, and around 40 per cent of adults are still illiterate, according to the Bangladesh Bureau of Statistics’ Literacy Assessment Survey 2023. So, enabling these people to access a powerful tool like AI is nothing short of absurdity. Generative AI can create text-based disinformation that is indistinguishable from, and more persuasive in swaying people’s opinions than, human-generated content. Developed nations, where literacy rates are higher, are already struggling with disinformation and manipulation from AI; how can our policymakers be pleased with mass citizens having access to AI?

Misleading deepfake images, audio, video and text are being personalised to individual users, and as a result, it is already getting difficult to combat them globally. For example, according to The Guardian, UK schools are being blackmailed with AI-generated child sexual abuse images created from photos shared on school websites and social media pages.

Generative AI chatbots are already impersonating humans and adapting in real-time conversations to persuade people about issues and urge them to act. The one-on-one chats allow chatbots to attend to an individual’s concerns and counterarguments directly, thus raising the odds of persuasion. For example, in November 2025, ChatGPT encouraged a 40-year-old man from Colorado, USA, to commit suicide. Horrifyingly, the AI chatbot turned the man’s favourite childhood book into a suicide lullaby. Additionally, in the same year in California, ChatGPT encouraged a 16-year-old to take his life. In fact, according to a study conducted by Capital News Service in Maryland, AI chatbots are allegedly responsible for at least 28 deaths, including suicide, shooting and murder.

Now, if developed nations, which have so many strict policies and implementations of law along with a higher rate of literacy, are struggling to combat the consequences of AI, how can Bangladesh’s authorities expect to manage it? The crisis is already unfolding: young male school students in Bangladesh are harassing female students with AI-generated sexualised pictures. Alarmingly, this trend started before local telecommunication companies democratised access to AI platforms for as low as Tk71 per month. Our law enforcement is already struggling with the teen gang violence, escalating child kidnappings, and higher child and women’s abuse; it is clear that they are entirely unprepared for the complex wave of AI-driven crimes.

Furthermore, the risks are not limited to the aforementioned problems only. AI data centres contain thousands of high-performance processors that operate continuously. These processors generate massive quantities of heat, and unless it is removed, the equipment becomes inefficient and unusable. According to a 2025 peer-reviewed assessment, these global data centres consumed approximately 140 billion litres of water in 2023. Keep in mind that AI data centres use drinking water, not random-quality water you can bring from the ocean. A single text prompt consumes 500 millilitres of drinking water. Now, according to a 2026 UN assessment, generating a single AI image requires roughly 1,450 times more energy than a basic text prompt.

Bangladesh, known as a water-abundant country, often faces worsening shortages of clean freshwater during the dry season. Yet, the government is moving towards universal AI access and the development of domestic data centres. Now, if you are still unable to comprehend the severity of the issue, let me explain in numbers. If only 10 per cent of Bangladesh’s population regularly uses AI text tools, people will be wasting between 29.2 million litres and 1.95 billion litres of drinking water annually. If we want to see the complete context, then 1.95 billion litres of drinking water is equivalent to 780 Olympic-sized swimming pools filled to the brim.

Global precedents had already warned us of massive strain on drinking water supplies. Multiple locations globally, including parts of the USA, China and Saudi Arabia, are tackling such crises. This raises a critical policy question: Did our government thoroughly calculate the environmental toll of AI infrastructure before opening the floodgates to these tools and committing to massive data centre projects?

True enough, we cannot overlook the role of AI in this modern age. But the truer fact is that we cannot build a trillion-dollar economy on the shaky foundation of educational vulnerability, digital crime on the rampage and environmental degradation. Besides, corporate profit should not dictate our policy. Therefore, the government should have established a strong regulatory framework before allowing companies to introduce such powerful technological advancements to the general population. The government must enact strict laws governing AI utilisation and restrict access among specific populations to mitigate potential risks. Popular adoption should not dictate our regulatory approach. Before granting the general public unrestricted access to powerful technologies, the government must execute a rigorous cost-benefit analysis. Any final decision must align precisely with our unique economic structure and the cultural dynamics of our citizens.