Is artificial intelligence a threat or salvation for humanity?

On March 4th, we held the GPT Party with 400+ participants, during which there was a fascinating panel discussion on the topic of artificial intelligence. In this and subsequent articles, we will talk about the most interesting moments of this event.

At the beginning of the meeting, Tair Mamedov threw us into a whirlpool of thoughts by asking whether AI is a threat to humanity or, on the contrary, its salvation.

Note: The speakers’ speeches are not presented verbatim. ChatGPT transcribed them into a clearer and more understandable text form.

Nikolay Davydov: “There is an obvious risk associated with automation for Knowledge Workers. When you produce robots and automate work with them, you have to manufacture, sell, and distribute them. It takes a lot of time. But when ChatGPT appears, it attracts 100 million users in 2.5 months. It sounds tempting, but for some reason, it’s scary. Anyone can start using it without special knowledge and skills, but to get real value from it, you need to spend time on it. And we already see that corporations forbid the use of ChatGPT, but engineers still use its codes because it’s easy and convenient. However, such rapid spread of ChatGPT will lead to companies and countries that use it to dominate the market. It’s like the appearance of chainsaws in 1905, when they instantly began to dominate the sawmill market. And what happened to the price of wood? Any sharp economic changes end badly because they throw many people overboard. But there are positive aspects too. The only thing that can save a person from being fired because of ChatGPT is their motivation.”

Nikolay Oreshkin: “The main problem with ChatGPT 3.5 is that it makes mistakes, and that’s only half the problem. The real problem is that it does so with great confidence. If you ask it a question to which it does not know the answer, it can generate a response that does not correspond to the request at all, but at the same time will sound so convincing that it can confuse anyone. For example, ask “Who is Lenin?” – and ChatGPT can answer: “Joseph Vissarionovich Lenin was born in such and such a year and did this and that.” Such confident behavior can lead to incorrect conclusions and erroneous decisions in the future.”

Albert Golukhov: “If we talk about AI and its risks for business, they can be divided into several categories. Firstly, there is the possibility of a complete change in the labor market. About 80% of businesses in America and other developed countries provide services that include content creation. If the cost of creating content becomes zero, the number of content providers can be reduced by orders of magnitude. This is a systemic risk that can affect entire countries. I won’t delve into conspiracy theory, but perhaps COVID-19 was such a test to understand how society would react to such a situation. This is one of the systemic risks that needs serious attention. The second systemic risk is at the level of military technology. Many countries are already creating AI systems that can control military actions. This is a serious systemic risk that can threaten people’s lives. Another risk is the creation of AI for the production of poisons and drugs. This can be dangerous, as people can gain access to high technology. And finally, bloggers may face a change in attitude towards content, as people may simply use AI to get answers to their questions instead of subscribing to a blog. In short, AI has its risks that need to be taken into account.”

Katerina Langold: “The speed of change is very high now. Our brain is designed to handle short-term stress well. However, when a wave of changes hits us and the reference points, so to speak, disappear, it becomes very difficult. I am concerned that the speed of change we will face in the coming years far exceeds the speed of adaptation of the brain, even for very talented and informed people. This is not because they won’t be able to understand or learn how to use the tools. It’s like going to a new store to connect a phone and buy a SIM card for those who have never done it. It may seem easy, but micro-changes create strong destabilization. We are already seeing a mental health crisis caused by this speed of change. I am interested in how AI can help our brain adapt to these changes. We need to take care of ourselves twice as much.”

Alexander Soroka: “Personally, my opinion is that if something destroys the world in the near future, it will not be artificial intelligence, but rather humans. Let’s look at the present time: I am much more skeptical about people than about AI. In addition, many people may lose their jobs in the future, and the government may start printing money to distribute to the population. Technologies continue to develop, this is not the first technological progress. Nevertheless, I am still optimistic enough and think that the future will be better. Perhaps we will lie on the beach and let AI work.

What am I worried about, honestly? The fact is that today most people absorb reality from the TV or some other sources. If ChatGPT becomes the main source of learning and information, then reality will be controlled at some point, again, by some large structure, one person, or perhaps a state. And it will pose the same danger that occurred in Russia today, where the main media is controlled, and the people who live there perceive a distorted reality today. And the same thing can happen in our life. I came in and said: tell me how many unicorns there are in each country? And ChatGPT gave me the layout there: in America, India, China, and so on. It turned out to be about 150%. I say, “Well, you probably made a mistake.” ChatGPT: “Yes, yes, yes, I’ll recalculate now.” He calculated it and got 120%. This is his reality. And a person who may never have studied mathematics will take it seriously in principle. Do you understand what we can face tomorrow? I think we are at the beginning of a long journey. What we have seen now are the first swallows. Today, he cannot analyze. He can generate, compose, but this is already a lot and impressive in principle. But I think tomorrow other waves will come, he will become stronger, he will learn to analyze, and managers will not be needed.”

Nelly Anderson: “Human communication is inherent in our nature. Do you remember when we were stuck at home for two weeks, a month? We all went crazy because we were like animals, we wanted to talk to each other. We didn’t want to watch Zoom, we didn’t want to watch YouTube, Instagram. Please, human, please. Thank God those times are over. I hope those times will never come back to us. It’s all about human communication. And, in fact, AI will probably only help. The question of blogging and media presence, I think that such threats, maybe in the near future, will be a story that you need to prove that you are not a deepfake. That is, to verify that you are a real person.”

Marina Mogilko: “A very interesting opportunity is to take my videos, analyze all my accents, jokes, and so on. And then say, let’s release a couple dozen deepfake videos. Every blogger has the task of asking themselves, if I stop filming, then what? We will create our own companies, I don’t know, come up with something. But in 10 years. Because we still want a personality that ChatGPT doesn’t have right now. And, by the way, one of the things that scares me is, for example, I want to hire a copywriter. And what, now everyone will send me text through ChatGPT? You’ll ask for a text about the English language, and then you’ll replace English language with a soft couch, and it will be the same meaning. As if such a stupid copywriter wrote it. The point is that you can write 15 idioms in English – great, thanks. But he can’t write a deep text about some really important issue.”

To be continued…