Interview by HATİCE NUR DICLE Artificial intelligence is an unknown field that many people are curious about. We talked to Anıl ÖZÖRNEK, who is an Al Evangelist on the TRAI website, whose artificial intelligence is sometimes very confusing and sometimes scary. WHAT IS ARTIFICIAL INTELLIGENCE?
Interview with Hatice NUR DICLE
Artificial intelligence is an unknown field that many people are curious about. We talked to Anıl ÖZÖRNEK, who is an Al Evangelist on the TRAI website, whose artificial intelligence is sometimes very confusing and sometimes scary.
WHAT IS ARTIFICIAL INTELLIGENCE? HOW DO YOU DEFINE ARTIFICIAL INTELLIGENCE?
First of all, I can say this: The concept we call artificial intelligence is actually a technology that is realized by the creation of certain algorithms at the very beginning and feeding these algorithms with the existing data. Well, how do I define artificial intelligence, I can say that it is algorithms that can do what people do and enable them to do what they do better. Here, too, the subject can progress differently because sometimes there are algorithms that can do all the things people do, and sometimes only algorithms make a job do better.
Here, “better” is actually better than the value proposition we give to the artificial intelligence algorithm. In other words, do we want it to be faster or do we want it to work in a larger direction. Now there is a situation. On the other hand, I will mix the concepts a little bit, but artificial intelligence actually consists of three different levels. So it’s actually a little bit of “what is artificial intelligence?” The question is confusing.
What I’m talking about is “Artificial Narrow Intelligence.” This actually allows you to do certain tasks or helps you do them better. Let me quickly give a few examples of this. There is an Algorithm called “Jipititri”. This is a natural language processing algorithm and thanks to the given data, this Algorithm can be found in an article or poetry can write. When we look at it again, for example, it provides a more accurate pricing in algorithms used in the private sector.
I’m throwing how much TL should this product be shown to Hatice, how much TL should be shown to Anıl. It does this much better than humans. The next level of these artificial intelligences is artificial intelligence, which we call general artificial intelligence at the human level. Another is super Artificial Intelligence. It is artificial intelligence that has reached superhuman abilities.
WHEN WILL WE GO TO STEP TWO?
There are two questions about this second AI. Can we go first? Most experts say we can go now. The second question is when can we go? It seems to take a really long time. So not anytime soon.
DO YOU THINK WE CAN TAKE STAGE TWO?
I think so too, but I also think that this is very important there. We, as humans, are moving forward with what we know. But there is so much we don’t know. Therefore, one of the limitations of artificial intelligence here is that we can get to that point as our understanding of how the brain works and our knowledge in understanding the functioning of a mental function in general increases.
SO CAN WE SAY SHORTLY: ACTUALLY, HOW MUCH MAN CAN UNDERSTAND THEIR SELF, CAN WE CREATE SOMETHING AT A GOOD LEVEL?
Exactly. For example, let me give an example from this last period. When creating artificial intelligence, it was always thought that cells were formed by considering a process in the brain. So, for example, I am speaking, you hear it, you form an opinion accordingly and you ask me a question. However, a recent study has shown that in fact, about eighty percent of the activities in our brains occur without any direct external stimuli, but for example, we haven’t figured it out yet. We don’t know what it means for the formation of cells.
WHAT MAKES ARTIFICIAL INTELLIGENCE TO US SO MUCH?
The benefits of artificial intelligence can be seen very clearly. That’s why we actually adopt it so easily and quickly. For example, they have small touches on our personal lives. But when we look at companies, for example, you are a retail company and on which shelf you ask “Which product will be sold more?” you think. This is a very difficult question for you. So it takes a lot of time for you to test which shelf to put a chocolate on. However, artificial intelligence can actually watch it with cameras and learn and analyze who buys it in what order, why it takes it, from which shelf, how it takes it.
Then you decide which product should stand on which shelf and how to stand. If you usually try to analyze this, it may take weeks or even months. This gives companies huge profits. or a clearer example of this banks To credit How much credit should I give? People used to do this, and for example, they were trying to decide on your past loans based on your financial situation.
Now, artificial intelligence is used in most banks. This provides a much greater benefit both in realizing whether customers will pay the loan they receive and in efficiency. Companies want to use artificial intelligence a lot right now. Because they know that the more they use, the more they will benefit.
IS NATURAL INTELLIGENCE SUPERIOR TO ARTIFICIAL INTELLIGENCE? WHY IS THAT?
Yes for today. Because today we do not have a more comprehensive artificial intelligence. Our current artificial intelligence cannot learn without learning something, without experience, without receiving that data. But we expect to see artificial intelligence at the human level, perhaps even superhuman, in the future. HE IS
there will never be such a thing. In fact, when you look at it now, some artificial intelligences can already make predictions, but they cannot do it without only experience, that is, data. In other words, it learns first, takes the experience and then makes predictions. People don’t need prior experience
for it to do. That’s why it’s superior right now.
DO YOU BELIEVE THAT THERE WILL BE A SOLUTION TO DISEASES WITHOUT ARTIFICIAL INTELLIGENCE TREATMENT?
Even now, I think these works have started. The company Deepmind most recently entered a protein folding competition. This company is a very, very good company in artificial neural networks. One of the most important events of proteins is that they always fold differently. For example, there are 10 protein series
11 how to fold. This is very difficult to predict. This is a very, very difficult problem and this problem could not be solved for 26 years. In the competition last 2020, Deepmind solved this problem with incredible accuracy. Once that was resolved, the scientists said, since viruses are also a protein, they’re right there.
They took the algorithm and used it for Covid. It has been of great help to the work there. In other words, vaccine studies have progressed very differently in understanding the virus. When we look at it like this, it seems to me that artificial intelligence will play a helpful role in finding the cure for diseases that we have not foreseen and that we cannot find a cure for now.
If you ask how artificial intelligence will provide this, in fact, artificial intelligence will provide prediction here. For example, artificial intelligence will take the data we have and will predict which of the 15 treatments is the most appropriate.
WHY CAN’T WE GO TO STAGE TWO TODAY?
There are actually many reasons behind this. First of all, our capacity as technology is still limited. So we can’t hold infinite data, we don’t have infinite data anyway. In addition, the machines on which we run artificial intelligence algorithms are still not at incredibly high performance. This is the technology side of it. In other words, the more data we can store, the cheaper and more efficient systems we can create, the technology part will come. Another problem is that we still have certain fundamental problems with understanding what intelligence is. We don’t know how to create. Besides, like other technological advances, we discover something but don’t know how to use it.
ARE THEORIES THAT SAY ARTIFICIAL INTELLIGENCES WILL TAKEOVER THE WORLD JUST THE THEORY? ARE THERE SITUATIONS WHERE ARTIFICIAL INTELLIGENCE IS OUT OF CONTROL IN THE FUTURE?
The starting point of this topic is the Terminator movie service, which both attracts the attention of artificial intelligence and creates fear from artificial intelligence. There is actually such a perception because of the Terminator movie. A lot of conspiracy theories have been produced from here. We think of conspiracy theories
We don’t really think about the other side. In other words, we do not think of a world optimized by artificial intelligence and where people live comfortably. There is an inequality here. So there’s more of a shift to the dystopian side. Secondly, is there such a thing? Yes, maybe.
Anything can happen. Actually, I think the issue that needs attention here is ethics. The issue of how we develop artificial intelligence is very important. In other words, artificial intelligence cannot take over humanity by thinking this. “I’m wiser now, I don’t need people.” Artificial intelligence is the value proposition we give it
whatever it does. Here, for example, the example of “Patchworks Factory” is used a lot. That’s it: you give paperclip data to an AI as a single value proposition, and the AI makes paperclips from all the metals in a factory. When the metals inside the factory are finished,
He continues to make paper clips with metals. This artificial intelligence does not stop the production of paper clips constantly. Now whose fault is it here? Artificial intelligence should be given the right value proposition here. In other words, value suggestions should be given to artificial intelligence such as “it should stop when it comes to this point, it should produce this much Paperclip”.
Goodwill is the key here. Discussions have already begun on this. In fact, I think it has been going on very intensely for like five years. Related to this, some companies, some investors established certain institutions to control artificial intelligence, produce it better and produce it in good faith.
As I said, the most important issue here is ethics. I can give an example for this. Artificial intelligence defeated one of the most famous Go players in South Korea. defeated 4-1. Nobody expected.
South Korea almost declared national mourning. The artificial intelligence company then realized that. We did our best, but it is not useful to people. Because the Go player stopped playing Go. For example, how is that artificial intelligence company artificial intelligence that makes people happier and does not harm people?
He works so that we can produce intelligence. There are companies that work with good intentions at the moment, but if one day a bad employee appears in a company, we will fight against it as a world.
IF ARTIFICIAL INTELLIGENCE WORKS WITH ELECTRICITY, WON’T WE TURN ALL ARTIFICIAL INTELLIGENCES OFF WHEN WE SHUT THE ELECTRICITY?
Yes, at the end of the day, everything is connected to electricity. Artificial intelligence will also depend on electricity. That’s right, but on the internet side, I’m not sure how possible it is to turn off the internet today. In the future, it will probably be able to work in artificial intelligence cloud systems. So this
This means that wherever there is internet, there will be artificial intelligence. Turning off the electricity at that point may not work either. It would be very difficult to close it, but I can say that there are servers that provide internet. These are huge servers. All of these can be turned off. Suppose there comes a time when
We said let’s close. Well then yes, it might work here.
WHICH OCCUPATIONAL GROUPS WILL BE ELIMINATED WITH THE DEVELOPMENT OF ARTIFICIAL INTELLIGENCE AND WHICH OCCUPATIONAL GROUPS WILL NOT BE AFFECTED?
Accounting-financial jobs with a very high percentage, and other automatable business lines are the occupational groups that I think will not end but that I think they can decrease relatively. In my opinion, the profession that will be least affected is teaching. But I don’t think that even a job will be needed at that point anymore when we move to the third level superintelligences. Maybe work can now be done like a hobby. Because when you want to produce something, artificial intelligence will be able to do it anyway. You can only do it as a hobby if you want,
WHAT WOULD YOU RECOMMEND TO THOSE STARTING IN ARTIFICIAL INTELLIGENCE RESEARCH?
I will suggest this for people who will not study about artificial intelligence, but who are curious about artificial intelligence. Which departments are otome in their own profession, what can be accelerated. Thinking about them, self-development and research on this subject is what I can recommend for this group. So these people should always research and see what can happen. Even the short “what is artificial intelligence” on YouTube like this? It would be beneficial for this group to watch videos that can understand artificial intelligence without going too deep. Another is someone who wants to work in the field of artificial intelligence, who wants to enter this job.
friends. First of all, I think they have to decide. Artificial intelligence consists of many areas, which of these areas is more prone to it. I recommend them to practice plenty of courses from udemy and similar sites for the field they decide later. Practice is very important here.
Source: Space Mag Turkey