One of the facts about AI is that it is mainly based around existing text and it is going to be a problem where the inputs are faulty. It wants to be a pretend human, and some of those aren’t good role models. I think we are going to see some failures. I was thinking about an application I put into government and the responses were a mess. Anyone who thinks that they will build an AI based on all the previous applications is going to have a big shock, it is going to get them wrong at the same rate as humans. They will need to go through a large sample of applications an determine if the response was correct.
1 Like
Not directly related to the subject of AI, but I had access to a wonderful book while working ten years in the world of law review publishing. The Editor-in-Chief had a book on his desk entitled “Words Into Type” that I just can’t say too much about or praise highly enough. You can look up virtually ANYTHING in that book - whether it’s the use of a word, phrase, punctuation, or grammar. It also has an excellent index in the back. Highly recommended!
2 Likes