On March 4th, we held the GPT Party for over 400 participants, during which an exciting panel discussion on artificial intelligence took place. In this article, we continue to share the most interesting moments of the event. You can read the first article here and the second one here. Note: the speeches of the speakers are not quoted verbatim. ChatGPT transcribed them into a more clear and understandable text format.
Ivan Pupyrеv: Once, me and the guys I work with went for a ride in a self-driving car. It cannot and should not violate the rules. Then we saw a van from which the guys were carrying a sofa. Our car stopped as it sees movement ahead. After a while, it realizes that people can move the sofa for a long time, so it violates the rule and drives around the van. It takes some time. At the same time, a taxi driver could violate the rule and have already driven around 20 times. I’m not downplaying the importance of artificial intelligence. This machine can do amazing things, and they are much safer than drivers. But artificial intelligence cannot do what it has not been taught. We teach it. Therefore, all problems are related to what we teach it. It is very difficult to imagine that tomorrow X-ray machines that photograph a person’s internal organs will give the wrong diagnosis and kill him. I think it is difficult to imagine that artificial intelligence poses any threat. Artificial intelligence is a tool. We create tools that become more sophisticated, and they do what we can do. But they cannot do what we cannot do.
Katerina Lengold: I want to bring together a few points because I think one of the comments that Kolya made is very important. Humans are amazing creatures who have achieved success because we have learned to invent various tools. We may move poorly and quickly, but we can build ourselves a bicycle. We should learn not how to make tools, but how to interact with each other very well and flexibly. Why did we win the evolutionary battle? Because we can unite in groups. Because we can interact brilliantly. Our brain works much more efficiently when we work together. When we sit next to someone and they tell us something emotional, we don’t necessarily have to say something useful in response. My brain will be drawing a picture of what the person is feeling. And when they feel that I understand their emotions, they calm down. Our interaction makes us human. I am very concerned that as we increasingly rely on technology, we communicate less with each other. I know Marina well. It is not what she says, but how she says it that matters. Sometimes I see things in Marina’s blog that I have seemingly seen, heard, and read millions of times in books. But how Marina talks about it, how she expresses her emotions, penetrates me much more strongly than all the rational advice. Therefore, content is not just idioms or a correct list, it’s about how a person talks about it with love for the language, how they admire these idioms, and how they can help us become a person of the world. It’s very cool! And that’s what makes us human, what makes us interesting. I really want us not to lose this, to remain more efficient, productive, and successful.
Alexander Soroka: I would like to ask you all a philosophical question. If we assume that humanity was created by someone (as many believe), then we, like artificial intelligence, were given free will. Therefore, should we give the same free will to the artificial intelligence that we create? Question from the audience: We talked about how artificial intelligence takes away jobs, but what new jobs will emerge, especially with the emergence of GPT, and will they become full-fledged professions? Albert Golukhov: Let me give you an example. There used to be a profession called radiology technician. This specialist took X-rays and then looked at the films to determine if there were any spots, etc. In the 90s, it was decided to make this process cheaper. India appeared, global globalization began, and files were digitized. Special scanners were created that were sent to India. There, low-skilled workers analyzed these X-ray images and sent the diagnosis back. Their accuracy was about 75-80% because a person could not compare every pixel as they got tired. Programs that later appeared began to analyze X-ray images with an accuracy of up to 98%. And all these people in India were no longer needed in this field. I am sure that in the future people will trust expert systems more, which have a matrix of 10,000 parameters, than some doctor who may be tired or not in shape. Most of us here are creative enthusiasts and entrepreneurs. However, most people are not like that. They are not capable of understanding some complex things. Those who have employees in their companies can support me. It is very difficult. I want to emphasize that unfortunately 80% of people are not capable of thinking complexly, and it is difficult to retrain them as astronauts, for example. To be continued…
