A (Limited) Defense of AI

Not only is this post late, but I am going to use it to say something that I guess is unpopular. I am going to defend ChatGPT. In a limited way.

See this post is late for a couple of reasons. First is that we have a Hotsy Totsy show this week and it is one with a lot of moving parts and it has taken more prep than usual. But that is still pretty much par for the course.

What is not par for the course is that I have decided to do a weird thing. I have decided to self publish a Science Fiction book. The book is called The Planet That Dreamt it was a Girl and I will be promoting it in future weeks, so I won’t get too much into the details of it here. But, it is near and dear to my heart, it is a story and a world and a character that I am really in love with. And with that, I have a lot of trepidations about going the self publishing route. The first one is that self -publishing has historically had a taint to it. That taint is less so now, but I am old enough to be queasy at the phrase. But I have also been shopping around a memoir to agents and publishers, going the traditional route, fighting all the old fights. I am pushing that book through the hellish tooth laden uterus of traditional publishing because it is a political book, it is a book about real lives, it’s timely and its about very hard things that I and other people I care about had to navigate. It feels necessary to try to give it the biggest, most well financed platform I can.

But my science fiction book. It’s innocent. It doesn’t deserve the abuse that a publishing push entails. It’s a sweet little story about a girl on a distant planet trying to deal with an alien and grief and being on a new world. I want it to keep being that for me and for whoever reads it. I don’t want to resent any part of the process of getting my little space pilgrim book out into the world.

That brings us to our defense of ChatGPT. See, another reason I have avoided self publishing anything is because, as I said in a recent post, I am a slow learner who does not follow a process well. When I see a list of steps and protocols and formatting instructions, my brain goes white and I just stop functioning. Someone would probably diagnose me with something based on that. They would probably say it was some kind of ADD or other neurological divergence and that is fine. Maybe it is. But really, my response to people when they diagnose me with anything via the internet is the same as it is when someone tries to sell me on a religion or a political movement or something they have just saddled with the word community, whether it warrants it or not - I do not want to join your group.

But whatever this is, Neuro-divergence or something having to do with the rain of semi-comic catastrophes that constituted my childhood, the end result is there’s a lot of things I don’t do because I do not feel confident that I can A. Follow the process to do it correctly and B. Ask someone questions over and over without ending a friendship or having to pay them to put up with it. 

But now we have ChatGPT. For someone like me, this is a miracle. I can ask, “What should the margins be.” And it tells me. Then when I don’t get it the first time, I can say, “Tell me again about the margins.” And it does. Then, when I don’t understand how to set the margins, I can say, “Explain to me where I find the margins thing.” Then when I fail to grasp that, I can say, “Explain it to me again, but this time, like I’m a fifth grader.” Then when it has rephrased it as though I am twelve, I can say, “And what numbers go in these fields again?” And it tells me. It never rolls its eyes, it never gets attitude, it never treats me like I’m an idiot, it just keeps giving me the information over and over until we eventually find a way that I am able to absorb it.  Then I say, “Ok, is all this true for the digital format as well or do we need to start over for that.” And it doesn’t sigh, it just starts over. It does not lose patience, get bored, or comment on how many times we have had to do this. I cannot tell you how that feels for me. After a lifetime of not asking for help because of the response I get from the people who might provide it, this is like a whole new world of doing complicated things is now open before me.

And yes, I know that you have to double check information you get from AI. And yes, I know that in spite of my saying that, people are going to comment on how I really need to double check information I get from AI. I know this because I know that other people are over confident in their reading and comprehensions skills and lack the clear eyed view of their own information absorbing limitations, than I am. Physician roll thine eyes at thyself.

So that is my limited defense of AI as an assistant to those of us who need things explained to us like we’re fifth graders. In the last couple of months, I have learned more practical and implementable things from ChatGPT than I did twelve years of public education. I will also forget 90% of what I have learned. But I can just ask the AI to explain it again. And it will. And it will not lose patience with me when it does.