AI Expert Dr. Parashuram Dahal Discusses Technology, Elections, and the Future of Work

Kathmandu. Dr. Parashuram Dahal works in the field of Artificial Intelligence (AI). He holds a Ph.D. in Physics and is an AI expert in the United States. He is at the forefront of technology, research, and innovation. Dahal, a former General Secretary of the Nepali Society in Texas, has made significant contributions to the education sector.

In an interview with Ratopati's 'Warpar' segment, Dr. Dahal discussed topics including the use of algorithms and AI in elections, whether AI can affect election results, its use and impact in elections in other countries, and the influence of Nepalis living abroad on Nepal's elections. Below is the edited excerpt of the interview (the full video can be watched and listened to):

• You completed your Ph.D. in Physics, teach, and conduct research in this field. What moment transformed you and led you to enter the field of AI?

My thesis was on trapping nitric oxide molecules with a magnetic field. I also used AI to calculate the probability density in that work. Because of that, it was easy for me to move into AI.

• If someone wants to know how AI works in simple language, how would you explain it?

The full form of AI is Artificial Intelligence. We humans have human or natural intelligence. Just like humans, AI is also trained by data. The data trains the algorithms, and that creates AI. The exact similar concept is to make it like the human brain, or smarter and more capable than that.

• What exactly is an algorithm?

They are tools used to solve a problem by applying theories of mathematical processes and using certain formulas.

• And what is AI?

It is about training a machine using processes to create artificial intelligence. Without algorithms, we cannot train AI data at all. There are correlations in that mathematical data, correlations among features. Algorithms are needed to understand what kind of correlations those features have and what the data patterns are.

• What kind of knowledge is required for a person to learn AI-related skills?

One must have good mathematical knowledge. At least a bachelor's level understanding of statistical mathematics and general mathematical knowledge. Similarly, one must have motivation and interest. One must be very hardworking.

• Do they also need to learn some programming languages for this?

Yes, they must learn. For example, Python, and it has advanced packages. There is something called Pandas. One must also learn to use various kinds of libraries.

• If someone says, 'I will get a job in AI and start earning from it,' how long does it generally take?

In developed countries like America, one can start at an entry-level position within three to six months. However, motivation, interest, and a strong background—meaning being a first-bencher in classes from school through college—are necessary from the beginning.

• How is AI currently working? Or, how does a person evaluate whether it is working or not?

There is an evaluation technique for that. Suppose there are 10,000 data points; we train a part of it, and then we evaluate whether the trained model performed well or not. We validate it. If the validation yields good accuracy and a good evaluation, it means the created algorithm worked well, resulting in good outcomes. If we get good results, we confirm that our AI model has performed well.

• There is a lot of debate about its reliability—how much to trust and how much not to. It has also created a kind of ethical risk. What do you say about this?

We see fake news and fake videos in the journalism sector. These are the current challenges. They are also the disadvantages of AI. Benefits and disadvantages always go hand-in-hand with everything, right? Use and misuse keep happening.

• Then how should one trust it?

If your coffee machine makes coffee perfectly, you have to trust it, right? If an AI-equipped robot digs your field, works in the garden, or cleans the kitchen, you have to trust it. This is the era now. There are 10-20 percent shortcomings.

• There is a common concern that AI will eliminate jobs and increase unemployment. Is that the case?

It is both true and not true. It has taken some jobs, yes. But it is not true because AI requires good minds to create it, to make AI even more advanced.

• What is the most advanced, astonishing thing you have seen AI do so far?

It is being used at its highest level in the current war between America, Israel, and Iran. Drones are being guided by AI, which is why bombs are hitting the exact intended locations. The highest level of AI is also being used in robots for space exploration.

• But didn't a bomb guided by AI hit a hospital?

Sometimes there are errors, about half the time.

• Let's not look many years ahead. Can we imagine what AI will look like in the next 10 years?

AI will become a part of our life. We might not function without it. A limited number of people will become smarter. But it will make some people lazy because AI and robots will do all the work. If robots do everything, people might become a bit more criminal-minded. There is a saying, an idle mind is the devil's workshop; criminal activity might increase. This is my personal estimation.

• There was a lot of discussion about algorithms in Nepal's recent election. The debate about algorithms winning or losing elections is ongoing. How do you view this?

The wave for the Bell symbol (Rastriya Swatantra Party) came spontaneously, driven by Gen Z youth, smart people, and first-benchers. They spontaneously created an algorithm in its favor.

• Can you create your own side using it?

Yes, you can. For example, if many videos supporting the Bell symbol are created, only those videos will keep appearing on the cell phones we use. This is just one reason. There was also disillusionment with the old parties.

• So, algorithms can significantly influence elections, right?

If that content keeps appearing every time you scroll, people get influenced. But someone is always working to make sure that content appears.

• The old parties give speeches saying the new party has no vision, no agenda, no policy, no destination, and questioning who operates them. Couldn't they have created that?

They were somewhat traditional. They had smart brains at the Gen Z level, but they lacked motivation. They were focused on money—where to earn money, which leader to follow for financial gain. Those who earn on their own, the smart ones, were not really in their favor. But people spontaneously supported the new party. The smart ones, the first-benchers, spontaneously created the algorithm.

• There is talk that Nepalis living abroad pressured their families to vote for the new party in the election. You also live in America. Did you apply pressure?

No matter the quality of work, Nepalis who have gone abroad command respect there. They have seen good governance there. They have seen the government there. They have seen how the police administration acts immediately and impartially when even a small mistake occurs, and how quickly things get done. People who have witnessed that experienced something different here.

There has been no improvement for 34-35 years; it has only been talk. RSP has educated people who immediately understand the sentiment. They have experienced these things themselves. They have studied both in the country and abroad. Therefore, the atmosphere was also set by those living abroad.

An UML leader was saying something like, 'How can those who fled abroad when the country was in trouble come back in the morning wiping tears from their eyes to criticize us? How can they criticize those who stayed to build the nation?' They say we fled; we say we were driven away, that we were not allowed to stay. This is because the last-benchers are in politics. The last-benchers threatened, looked down upon, and insulted the first-benchers, and did not let them stay. That is why the first-benchers chased the last-benchers away in this election. In my opinion, it will be very difficult for these last-benchers to recover.