AI is now everywhere, but no one has seen it. It can still already improve our lives

No items found.
19/4/2023

In Trask, we trust that progress is the main force of our work. Currently, the buzzword "AI, artificial intelligence" is a subject everyone is talking about. The problem is that almost no one really understands what it is. That is why we asked our experts from Trask, Andrej Svitek, and Milan Kubálek, to help you be smarter when someone talks about how millions of people will lose their job because of AI.

You might also enjoy

Read more

[.infobox]ChatGPT is a chatbot from OpenAI published in 2022. It can write computer codes, poetry or pass school exams. Although some its answers are not correct, its launch brought massive attention to AI in general.[.infobox]

We see the end of certain professions in the context of ChatGPT in the opposite way. "Thanks to similar technologies, we believe there will be a new profession, and many existing disciplines will transform. And then they will be better than before," our experts calm the situation.

And in fact, popular ChatGPT is not an example of artificial intelligence. The definition of general AI is that AI can learn itself and decide what it will learn. "ChatGPT is a potent tool that answers according to what people taught him, but it cannot create anything new. So, it is very far away from general artificial intelligence," Our experts say, although they agree that even ChatGPT can potentially change our lives.

It is like the arrival of electronic books. They did not replace classic paper books, which still have their place and value; their market share is smaller than it used to be. Similarly, a journalist can spend more time creatively, and ChatGPT can help them with research and analysis of generally available information because that will be significantly easier.

AI will improve people's performance

How do we know that artificial intelligence can improve our careers? Although AI is a new phenomenon, best example of its potential impact is 2,500 years old game Go. It is much more complicated than chess, so experts thought computers had no chance against human champions in it. And in 2016, Google did it. Its AlphaGo defeated the best players in a different style than everyone has played. "After humanity spent thousands of years improving our tactics, computers tell us that humans are completely wrong," said Ke Jie, the world's No.1 ranked player.

Google released its tool so players could learn with it, which is AI's most significant benefit for humankind. A study by Korean scientists discovered that Go players play much better than before AlphaGo. Interestingly, the best outcome was for senior players; juniors without such experience did not benefit from AI as much, although they have also trained with it.

Strengths and weaknesses of ChatGPT

So, one old Chinese game shows that even though we cannot be better than a computer, we can be better thanks to computers. In what activities are computers better than people even today, according to Trask experts?

– speed and quality of obtaining information – you cannot read so much in your whole life as ChatGPT already processed

– tireless – answers to anything 24/7

– synthesis – summary of a long book into one paragraph, so you do not have to read it

– repeating tasks – if you are bored creating new contracts, give data to ChatGPT, and it will do it for you

But ChatGPT is only start of the evolving AI. What are its problems? Trask experts warn against these things to be wary before you use it:

– no context – ChatGPT has no context about profound consequences, human emotions, and reactions to them

– prejudices – people are not saints, so the system, which has learned from people's data, is not better

– no verifying – the primary purpose of this chatbot is to talk, not to say true. So often, when ChatGPT does not know the answer, it will create it – people, books that never existed

– old data – ChatGPT has no access to the Internet, and its data ended in September 2021. Anything newer does not exist in its cyberspace

In conclusion, ChatGPT is unable to invent new things. It can only respond to topics that it has been trained on. It cannot think of anything new, and it is not creative. It is a very advanced and complicated model working with statistics. It can potentially influence how we will work in the future and how we will operate in general. But it is not artificial intelligence. It is not pure AI. And we understand it perfectly anyway.

Written by

No items found.