“This was not created by AI software,” says the AI software

Not so very long ago, there was a class of jobs that computers or robots could not take from human beings: the creative arts. Electronic brains, with or without mechanical appendages, could build cars, perhaps drive cars, calculate tax refunds, verify identities, cook food, sell products, explore planets, and harass motorists, but they could not write a book, act in a movie, or draw a picture.

Until now, at least to some approximation. Yes, in many cases, pictures of human beings produced by artificial intelligence still look like cartoons and have too many fingers, while their essays have a clumsy writing style and make basic factual and interpretative errors — not so different from a lot of human writers — and the Virtual Voice audiobooks are not too bad for a computer reader with less emotion than a partially anesthetized Robby, but all of this is a lot of progress, if we can call it that, for something that until recently was grinding along at the level of hopeful speculation and workshop tinkering.

The complaints from content creators of all media are to be expected and fall into three main categories: 1. The artificial intelligences must be trained, and that is done by showing the computer the work of others to copy, 2. The products are on the verge and in some cases have crossed over into being difficult to…