The “intelligence” of language learning models is a subject I have wanted to broach for awhile. As Artificial Intelligence becomes more impressive it seems the definition of “intelligence” in some circles is changing. I have seen videos on LinkedIn where people claim nearly every job will be taken by Artificial Intelligence, and honestly I disagree. Having worked with robots in manufacturing I have come to find they are great at repetitive tasks. That is because they do not learn in the same way we humans learn. Machine Learning and Artificial Intelligence rely on previous knowledge and statistics to make reasonable conclusions. That being said their conclusions are highly contextual. In other words, they cannot infer truths about other related things without being trained how to think about that thing. While general education classes in college are boring they teach you to avoid this very pitfall.
I have used ChatGPT and it is a very neat tool. I asked it for advice or content on a variety of subject matters, and I thought it did a good enough job providing a rough draft. That being said, it will not make anyone an expert in any given field. Without the sort of knowledge that comes from proper studying and experience I do not think the tool is very useful at all. Furthermore, the little errors it does make do not really improve the robot processes that people already dislike. I do not think that robots will become sentient or take your job unless it’s incredibly easy to automate.