AI is an impressive tool. It can tell almost everything, But sometimes it makes very big mistakes. I once asked about altocumulus, which is one type of cloud. The AI gave information about some kind of flower. The fact is that these kinds of errors are good examples of mistakes that AI can make.
The AI uses the page rank of the search engine. And that makes it easy to manipulate the answers that it gives. I use the AI just for fun, and in some cases, I use the sources that the AI gives. But there is a long journey to common AI that can create things like humans.
***********************************************************
When somebody asks something by using AI, that person gives an advertisement to that system.
When a user asks something, the AI must have certain instructions on how to search for data. If there are problems with queries, the AI cannot form the entire data set. And if the query is wrong, the AI is unable to operate.
The AI requires fuzzy logic for fixing writing errors. The problem with fuzzy logic is that if some words are similar to each other, like back, black, and slack, that causes a problem with fuzzy logic. The AI doesn't know if the user means black but writes "slack".
That's why queries must be long enough so that the AI can use other text to conclude what the word that person means. When humans fix those kinds of things, they see from context or entirety what the word means when the user writes the word black.
And that causes problems for AI in making answers. The AI doesn't give answers by using imagination. It just connects different data sources, and then it generates answers by connecting certain parts of those databases.
The system must search its internal databases. Then it must go to the open internet for search data. The problem with that process is that the AI has problems with cultural and language differences.
************************************************************
Originally, the idea of AI was to develop a system that could replace humans as ordinary or routine thinkers. The purpose of the AI was to outsource routine thoughts to it. And that is one of the things that people don't think about: they outsource their thoughts to the AI. And that thing is not a very good idea. Our brain requires practice, and we should not stop thinking even if we have some kind of AI.
The AI is in its infancy. That means there are many types of bugs in the AI. The thing is that the development of AI is just beginning. The AI is a suitable actor if it has a limited operational area and sector that it can effectively control. Making some new molecules is different from making them act like real humans.
So at this moment, the AI that discusses with people is suitable for fun. Or the AI can be used to make some kind of computer program because in programming there are certain, clear orders. That makes the AI very good when programmers want to make some database connections. But the problem with AI is that it doesn't know what something is. And in some uncommon cases, the AI is helpless because the search engine has not given a pagerank for that kind of topic.
Comments
Post a Comment