Librarians have heard about AI (Artificial Intelligence) but you and I as librarians may not know how it will be affecting our profession and jobs. Publishers and writers are worrying, because some of the books being written by machine are actually quite good. Whether this will effect how publishers deal with book submissions, or how libraries will catalog these things without an author, but a machine number, is up for wondering, as well as worrying.
I suppose there are lots of us writers, editors, and librarians who are going to like what a special AI machine can do to help with the load of getting the book written, on the shelf for reference or circulation along, with audio book making, so we don’t have to read everything, only listen.
Some of the issues being discussed relate mostly to publishers and writers; will anyone be left to write good stories? If too many books come out because of the swiftness of how many can be written, and published, where will they go, how do we sell them, review them, how do we make shelf space for them?
We now have many books about AI and the new 2022 best list can be found here.
A new Generative Pertained Transformer-3 (GPT-3) software is doing really well. It is a broadly useful language algorithm that utilizes machine learning to interpret text, answer questions, and accurately compose text. It analyzes a series of words, text, and other information then focuses on those examples to deliver a unique output as an article or a picture. Here is an article about its first book. The article also discusses some of the questions. If AI makes up a story from things already written, who owns the
copyright? GPT-3? Is this the ‘Author’?
We have just recently seen how AI has written a song, and recorded the song using an AI singer to sing it. How does a librarian cataloger, or document this? Doing a search I found there are AI song creators, generators, writers, remixers, et al. One AI song has caught many in wonderment. In just a few days it accrued over $330,000 in streaming revenue.
Would people using an AI book creator make as much money as a writer writing a novel? We do have a book which tell writers how to publish books in Public Domain, but republishing digitized content/words, would be quite different; perhaps plagiaristic. It seems, from reports, CPT-3, while having lots of buttons to push, still can’t make much of a story from what it has to work from.
Some things AI are encouraging and are already being used in libraries, but it appears AI isn’t going to take over the library book or record collections very soon. I have an essay writing software program which will search everything which has been written on a subject of choice; mostly on the internet. The program will also cite where a particular paragraph has come from. I’m not sure using those paragraphs and such would be very useful, if I’m trying to add to knowledge, rather than report it; I’m sure AI would have no idea what my review of a book would look like either, so I only use this program to see the many ramifications of a topic. I can also do this with the carrot2 program for AI in libraries.
I certainly believe we are going to learn of issues documenting, finding, citing, AI created content, along with legal issues of who owns what in AI content. A start would be to check out this article by Westlaw.