Topic: [Article] How ChatGPT Will Destabilize White-Collar Work (Read 797 times)

And I understand that the emergence of AI and its wonderful applications in work is really effective to help people run their lives more quickly, instead of being afraid that it will replace people in the workplace in some areas, learn how to use and control it.

~

Most companies have strict information secrecy rules put in place that prevents employees from sending confidential company data to 3rd-party servers. Since ChatGPT is trained on some of the questions you asked it, that means it could potentially leak intellectual property secrets, which is why places like Samsung have already banned its use internally - and also why white-collar jobs aren't in immediate danger ( the blue collar jobs though, I would be more worried for them as even automation can make them redundant).

We can point out AI's limitations and we cannot underestimate the capacity and capability of human to get the job done, the developers of these AIs' are not meant to replace human, they developed AI to be a tool to help human makes their job easier and simplified.

It just happens that AI works perfectly because humans aspire to create a perfect tool and because of that perfection, we think that AI will replace humans.

I don't want to think that we are going to be in Terminator-like scenarios, because AI is still limited, Thee AI cannot create AI they will never become creators of another AI and they should not be programmed to create another AI, because a human has insight that we called hunch or intuition while AI just relies on data.

I'm personally skeptical on the ability of machine learning/neural networks to fully replace humans. I think it's more likely that it will boost productivity by performing the more simple and routine tasks, but there will still be the need for human touch. People will lose jobs because now a professional will be able to do the work of 2 professionals, but AI won't do the work of all professionals.

- However, I'm not sure what can be allowed, for example, to engage in the complex development of a large project. My area is IT, and here he is, well, to put it mildly, not a "star" everywhere.

I'm personally skeptical on the ability of machine learning/neural networks to fully replace humans. I think it's more likely that it will boost productivity by performing the more simple and routine tasks, but there will still be the need for human touch.

The problem is that we often confuse learning systems/neural networks with artificial intelligence. These are completely different concepts and systems. And the goals are different.
ChatGPT can hardly be called an artificial intelligence or a "matrix" In fact, it is a huge repository of information that can find the right answers among trillions of blocks of information and formulate answers similar to human speech. A sort of Google search engine is a wrapped human-like interface for issuing search results.
Yes, neural networks and other technologies work here to collect and systematize primary information. But this system has no INTELLIGENCE.
I hope you have already seen this masterpiece: "The neural network was asked to draw a salmon swimming against the current. Well, she drew it."



In ChatGPT, the picture is much better, but the intelligence never appeared there. Since one of the areas I do is IT (development), and like everyone else I saw how ChatGPT "produces ready-made ideal code." I tried to set a task for him, something from the area "Write code so that when you click something, something opens somewhere." The system began to ask clarifying questions so that the problem statement was extremely accurately described in order to understand which "code from the knowledge base to take." If there is no clear statement of the problem, you will not receive any code. More precisely, if there is a place where there can be variability and you need to figure out the essence of the process yourself, ChatGPT will not help you. Therefore, if we talk about "replacement" - perhaps he will replace some of the "pioneers" of developers, who are essentially "translators" from human to programming language, not difficult tasks. Although I agree that as the system learns and the knowledge base accumulates, the level of "replacement" may rise

Yes, it seems the most likely outcome is that AI will boost efficiency of existing workers rather than fully replacing them. Which still means that hundreds of millions will lose jobs. And if tech companies will have high profit margins due to boosted productivity, maybe high taxation and wealth redistribution will become more viable than it is now. Plus I think some entirely new job opportunities will be emerging with new technologies and trends, so humanity is not doomed like some people like to say.

I have seen it writing near perfect code. When you don’t like a certain part, it modifies that part and comes  up with a better one. It is definitely not a copy/paste search engine.

In the end it is just a tool and can be as smart as its user.

When you tell it to write a poem, it will bring a crap generic poem. When you give it lots of details it will write a way more detailed and original poem.

AI has been a buzz word for long time, I still remember Deep Blue "beating" Kasparov and all. But I do not think that most white collar work can be achieved even in the next decade. It may be a close collaborator to white collars, basically removing the need to do the repetitive bits of the work. I am kind of happy that google guesses half of the time what I am going to write... but still is only 50%

AI has been a buzz word for long time, I still remember Deep Blue "beating" Kasparov and all. But I do not think that most white collar work can be achieved even in the next decade. It may be a close collaborator to white collars, basically removing the need to do the repetitive bits of the work. I am kind of happy that google guesses half of the time what I am going to write... but still is only 50%

https://www.theatlantic.com/ideas/archive/2023/01/chatgpt-ai-economy-automation-jobs/672767/


I'm personally skeptical on the ability of machine learning/neural networks to fully replace humans. I think it's more likely that it will boost productivity by performing the more simple and routine tasks, but there will still be the need for human touch. People will lose jobs because now a professional will be able to do the work of 2 professionals, but AI won't do the work of all professionals.