AI: A blessing, or an uncontrolled fire?

Shahria Islam Shaila

Artificial intelligence, or AI, brings hope on one hand and growing worry on the other. It has made many things easier, from detecting disease in medicine to analyzing weather and production in agriculture, simplifying work, supporting education and analyzing information. The real question is what happens if this powerful technology is misused. Even the world’s superpowers are worried about this and have begun to see serious risks in AI. For a developing country like Bangladesh, these risks deserve even more careful thought.

The biggest fear AI has raised concerns jobs. A 2025 World Bank analysis found that demand for AI related jobs is growing fast in South Asia. At the same time, demand for some easily replaceable jobs has fallen because of AI. This is a major warning for Bangladesh. A huge number of young people enter our job market every year. If AI takes over work in garments, banking, call centers, accounting, translation, data processing, content creation and software, the pressure of losing old jobs will grow alongside the creation of new ones. Those who can use the technology will move ahead, while those who cannot may fall behind.

The second major danger is fake information and fraud. AI can now fake a person’s photo, voice and video. This means someone’s name can be attached to statements or videos they never actually made. Once ordinary people find it hard to tell truth from falsehood, rumors, fraud, character assassination and social unrest can spread quickly. Signs of this have already appeared on social media platforms. International AI safety reports have already documented the risks of AI driven fraud, fake content and influence over people.

The impact could be even worse in Bangladesh, since we already know from experience how fast a rumor can spread on social media here. In the future, if thousands of people believe a fake video or voice recording is real, it could harm a family, an institution or even an entire community.

The third risk is cybercrime. AI can be used for good, but it can also become a powerful weapon in the hands of criminals. AI has raised the risk of fraudulent messages, exploiting human weaknesses, creating fake identities and preparing cyberattacks. International research has also pointed to evidence of AI driven cybercrime and its potential spread. If Bangladesh’s banks, mobile financial services, government databases and ordinary people’s personal information become more dependent on digital systems, security weaknesses could cause major damage. If someone’s bank details, identity, photos or personal documents are stolen, the damage will not stay limited to money. Their personal life could also be put at risk.

The fourth risk concerns the security of personal data. While using AI, people may end up sharing their personal information, business data, government documents or other sensitive details. Ordinary users often cannot know where this information is being stored, who is using it and what it may be used for in the future.

This raises an important question for a country like Bangladesh. Should we only use foreign AI, or should we also keep control over our own data and technology? Bangladesh’s draft national AI policy also mentions the need for effective national control over important data, infrastructure and AI systems through digital sovereignty.

The fifth risk involves national security. If a country’s official documents, critical infrastructure, economic data or security related information become unsafe through AI driven systems, the damage could be enormous. There is already international discussion about the risk that future AI driven cyberattacks could try to disable a country’s power, banking, communication or government systems.

An even bigger issue is the competition tied to war and power. The technological rivalry between the United States and China over AI is no longer just a business competition. It now involves questions of national security and strategic power. The two countries are set to hold bilateral talks on AI safety next week. This shows that the future control and safety of AI have become a major concern even for the world’s biggest powers. If AI is used too heavily in military decisions, intelligence analysis or autonomous weapons systems, the consequences of a single wrong piece of information or a wrong decision could be severe. The same technology that could help prevent war could also, if misused, increase the risk of war.

A worrying possibility is that if people start trusting AI decisions more than their own judgment, human ability to make independent decisions could weaken. Experts still disagree on how controllable more powerful AI will remain in the future. However, international AI safety reports have specifically highlighted the risks to human control and autonomy.

The example of fire fits best here. Fire is something people need, but it brings terrible destruction once it goes out of control. AI is much the same. The answer is not to put out the fire, but to create rules for using it safely. AI needs the same kind of control system.

Bangladesh’s first task should be to set rules and guidelines for high risk uses of AI. Safety checks must be made mandatory before AI is used in health, law enforcement, banking, national security or important government decisions. Second, the security of personal and state data must be ensured. Third, technology and laws must be developed to detect AI generated fake images, videos and voices. Fourth, there must be independent oversight and accountability for the work of AI companies. Bangladesh’s draft national AI policy already proposes risk assessment, documentation, transparency and oversight for high risk AI systems.

At the same time, people also need to be made aware and cautious. No one should believe a video the instant they see it, or accept a voice as real the moment they hear it. Future citizens will need to know not just how to use technology, but how to recognize its deception.

AI cannot and should not be stopped. What is needed is to use it within limits. Bangladesh’s goal should be to build a system where AI makes people’s work easier without leaving their jobs, data, security and independent decision making unprotected. Still, whether AI is good or bad is not such a simple question. The real question is who holds AI’s power, under what rules it is used, and who takes responsibility when something goes wrong.

We keep fire in our homes, but we never let fire become the owner of the house. The same principle applies to AI. Technology should remain under our control, and it should never end up controlling us. That should be the core idea behind Bangladesh’s future AI policy. The most responsible path forward is neither fear of AI nor blind faith in it, but using AI with safety, control and human interest placed first.