Future Job Replacement by AI?

The idea that AI will massively replace human jobs in the next several decades seems to be widely accepted. But here's the thing, lots of things about human beings haven't changed in decades, so why would we ever expect a mass shift in human employment due to AI?<br><br>One example of this is for the past 30 years, people have been trying to create a sentient AI by combining lots of large language models together. And for the past 30 years we have been failing spectacularly at doing so. I could see us having a key break through in this area within the next decade, but why would that matter? Many people thought about even basic models like Siri as speaking human beings who lacked the ability to say the right thing, but within a few years there was a very large divide between how humans spoke and how AI models spoke. And that's just the difference, with more or less advanced models, humans just like to speak differently than AI.<br><br>Even if we get to the point where AI models can do the same things as humans, we will still need to have humans doing those jobs. For instance, if an AI model was trained on the collective knowledge of humanity, there would be a huge problem with bias. We've all be told to never trust a Wikipedia article, because even though it was created in the spirit of crowd sourced knowledge, it may not be entirely the truth. So if you have an AI model trained on the collective knowledge of humanity, it would be based on the collective biases of humanity. At the very least we would need to have humans working to keep those biases out.<br><br>In many industries, like for instance medicine, AI models may never replace humans doing the jobs. Even if we had a model that could diagnose or treat illnesses with 100% accuracy, would we even want that? In medicine there is often a huge bias towards certain treatments, they often seem to not treat illnesses by trying to get to the root cause of them, but just trying to get rid of them. Maybe this isn't true in all cases, but it's definitely true in many cases. So what if an AI model was able to diagnose or treat illnesses with 100% accuracy? A lot of the time it may just prescribe some medication, which is just a Band-Aid on the root cause of the problem.<br><br>In other jobs, like for instance a hospital, we will still need humans. One example of this is in cleaning. Right now, hospitals are cleaned by humans, but what if we replaced a lot of the jobs with robots? At first glance it may seem like a great idea, but what if all of the robots were able to get a virus or something. For a human to get a virus and have it get passed around, a sick person must take a shift and pass it to everyone they come in contact with. But now imagine that every cleaning robot is infected with a disease. All they need to do it walk down a hallway, and all of the other robots may be infected with that disease. This is just one example of why we may not want to replace all jobs with robots.