In the current era, artificial intelligence or AI is being portrayed as the key to the future. Governments, industry, academia and the media have repeatedly said that AI will solve every problem. Medicine will improve. It will be taught in schools, it will bring fairness to the courts. It will bring revolution in every sector including agriculture, banking, transportation.

Today the word AI has become a 'trending term'. At technology conferences, startup pitches or policy speeches, AI is portrayed as a 'magic wand'. It gives the impression that it will solve all the problems of human life. Its use is claimed in every field like health, education, agriculture, employment, etc., but most of them are based on only theoretical attraction, without basic studies. Such claims push society beyond the real challenge . So that we can fall into confusion . This trend has a more dangerous impact on developing countries like Nepal. Where hype is more trusted than technology testing .

Is AI really a reliable tool for decision-making ? Most of the AI ​​systems in use today are 'predictive', i.e. those that try to make predictions, but most of those predictions are flawed . These systems are based on old data and patterns. Who cannot grasp the complexity of the present society . For example, an AI system that measures a person's likelihood of committing a crime, which can make judgments based on historical data that are racially or socially biased. AI cannot answer the questions of why and how, it only guesses what happened . Therefore, it should be used very carefully in the decision-making process. The use and capacity of a system that does not have a proper data collection system like ours is still limited and misleading.  To break the story, AI is achieving remarkable success in the field of language translation, face recognition or object recognition. Such systems are particularly visible in Google Translate, face recognition apps or biometric tools based on machine learning. But their limitations are also evident, including an inability to understand cultural contexts, language differences and poor performance in underrepresented groups, and an inability to catch exceptions. For example, the current AI seems unable to accurately translate the meaning of a Nepali proverb or parable or specific idiom into English. It is clear from this, that AI systems are effective only in specific and controlled environments, but they have many limitations in open, complex and multidimensional life situations. 

AI technology began with pure scientific curiosity. But now its direction is determined by the market . Investors in AI development today are more investors than researchers. Those who look at the return on investment, are not true . According to one study, most current AI research is not replicable. Which means, their results are based on propaganda rather than scientific validation. In a study published in the journal Nature in 2016, 70 percent of scientists could not reproduce the research of others.

s are often written for headlines, not results. This trend is also increasing in Nepal. Projects in the name of AI are being implemented as programs without testing, data or impact analysis. When such programs form the basis of policy making, they can lead to serious confusion and long-term effects .

We need to understand, AI is no longer just a matter of technology . It's big business now . The world's biggest technology companies Google, Microsoft, Meta, Apple, OpenAI are all investing billions of dollars in AI. Their aim is to dominate the market rather than human services. Concepts like smart classroom, smart health, smart city are attractive, but often they are based on limited laboratory tested concepts. Those systems do not address the complexities of real life, poverty, ethnicity, language diversity and geographic context. When we policy such systems as they are, we are ensuring failure .

In developing countries like Nepal where resources, access and manpower are limited, it is a serious mistake to accept technology as superstition. For the success of any technology, it is important to understand the local context, behavior and needs. An AI made in a foreign laboratory cannot capture the multifaceted character, linguistic diversity or sociological differences of Nepali society . Policy making based on such systems can set wrong priorities . The price of which can prove to be expensive in the field of education, health, or employment . That decision cannot be revised and the victim is always the people.

Every day we hear in the news lately that AI has written a poem, created a picture or diagnosed a disease . Such news creates miraculous expectations in the general public towards technology . But how much of this news is based on facts and how much is propaganda? The media itself has become propaganda oriented without deep knowledge about the use and impact of technology . Instead of translating complex concepts into simple language, the style of 'this is the future' is also including journalism in the market game . Its impact is also seen in education, where students learn not to review technology, but only to use it .

is not a positive possibility . AI in Nepal can be useful in agriculture, health or administration .  Such as: insect identification, road accident analysis, primary disease diagnosis or government data analysis etc . But for such systems to be successful, three things are essential. First, an understanding of local data and context. Second, transparency and accountability. And thirdly, the thought of including all classes through the benefits of technology . If technology is concentrated only in cities or capitals, rural communities will be further left behind and unequal access to technology will increase social divisions .

Now is the time for us to begin the practice of reviewing AI technology without blindfold imitation. The foundation of science is questioning, so we should have the same approach to AI. From school to university, not just 'how to use ?', but 'why use ?' The practice of raising the question 'What will be its effect?' should be increased. For that, it is necessary to make education critical. Education should work towards making this generation not just a consumer of technology, but a critic.

Today we stand at a decisive moment, where the speed of technology is faster than our consciousness and faster than our ability to make decisions . AI is no longer just a tool, it has become the basis of our thinking, behavior and society. But is this technology really being used for our benefit? Or are we unknowingly becoming its slaves? The answer to this question is not easy, but very important. AI in itself is neither good nor bad . Only its use determines its value . If we make AI a mere tool for production, profit, and propaganda, it trumps humanity and inclusivity . We have to think . Can a technology be considered useful that does not listen to all sections of society, does not include the voices of minorities or respect the local context ?

Therefore, we must rethink our relationship with technology . We should be able to make AI a means of expansion of ideas . An idea that prioritizes inclusiveness, transparency, and accountability. Education should teach not only the use of technology, but also its ethical review. When making a decision, policy makers should not only look at innovation, but also think about its social impact, long-term consequences, and its relationship with ethics . Scientists and researchers should step out of the laboratory and listen to the voice of society. And the time has come for the common people not only to be consumers of technology, but to be conscious citizens and ask - is this technology making us free or controlled?

If we do not raise the right questions today, tomorrow's society will be very inconsistent, where decisions will be made not by people, but by systems that do not know them . If we want to make technology a servant, we need to develop the consciousness that can question, review and direct it from today . Because AI doesn't make decisions on its own, we do . Therefore, today's decision and tomorrow's future depends on our conscience, thinking and sensitivity. Science is only useful as long as it is consistent with humanity. So is AI . Otherwise, it can become just another kind of 'snake oil', one that dreams but loses consciousness .