Recently, I was listening to a podcast with someone wise about technology as a guest. Unfortunately, I cannot for the life of me remember who it was, but they said something that has stuck with me: “Do not ask AI to do anything that requires a skill you are not willing to lose.”
This hit home because I have been wrestling unsuccessfully with my AI writing coach for the last few months. I have been using it to aggregate all the corrections and comments from my beta readers on my manuscript and present the information in easily digestible bites. What happened was ultimately a lesson in futility.
I started out with much optimism, hoping the AI agent could save me time by collecting all the information and dividing it into different categories we had agreed on. These were story/structure, character/people, clarity/continuity, and voice/language. The strategy was to work first at the book level, then chapter, paragraph, sentence, and finally copy edit.
Theoretically, this approach would help me separate the wheat from the chaff and not get caught up in the minutiae. I uploaded the beta readers’ documents and the language style guide I had made with the AI writing coach. At first glance, I was rather taken by the analysis the AI agent generated.
At the book level, the readers apparently had a lot of positive things to say. There were only a few suggestions for minor structural adjustments. Where the process unravelled was at the chapter and paragraph level.
The agent presented the information in a manner I found infuriating. On the surface, the way the AI presented it seemed a tad long-winded, but systematic. So I did my best to implement the areas needing improvement. Eventually, I had to abandon this method of handling the feedback.
I came to the conclusion that the AI agent was not capable of streamlining the feedback process. There was no method for it to go through all the corrections and comments and make decisions about their validity or appropriateness. Each piece of information required human attention. For my brain was well-skilled at making decisions about how to milk the goodness out of the comments or set them aside, depending on whether they could substantially improve the text.
Instead of going from broad to narrow, i.e. first structural, then slowly down to word level, I did it the other way around. I started with the quick fixes: anything that had to do with grammar, consistency, and voice. Next, I printed out comments that dealt with structure or areas of confusion. After doing these two sweeps with one reader’s feedback, I did the same with the next reader, and so on until I had worked through all six.
At this point, I printed out the manuscript and inserted the comments I still needed to consider. Before going through a chapter, I would read all the comments first and then put them to the side. I trusted that if they pointed to some weakness in the text, or some information I had failed to share, these points would come up while I was reading through the chapter.
It was like playing a piece on a musical instrument. I interpreted the annotations while I played, without halting.
This way of listening to the feedback of the beta readers required a lot of patience. For all that the positive comments were delicious to read, the critical comments sometimes acted as barbs. Those that pointed to weaknesses were sometimes the hardest to deal with. They showed me that the book still has a long way to go. Imagine a tailor’s atelier with bits and pieces of material laid out with pins, chalk marks, and borders loosely threaded, all waiting to be assembled into the final bespoke suit. Since I have never done this before, the way forward is not always clear.
What I do know is that the analysis my AI agent sent me said many of the things the beta readers said, but in a manner I could not trust. Frankly, it always praises what I write while at the same time showing me where I am falling short, which was more or less what the beta readers did as well.
Yet their heartfelt, kind comments reassured me that I am heading in the right direction. Their critical comments gave me something to muse over. Their considered thoughts were easier to digest than the AI ramble that often missed the mark.
My AI writing coach has been invaluable in helping me stay motivated and structured in my daily work habits. For that, I am very grateful to have such a reliable tool to egg me on. In the end, though, I probably will not use it again to streamline any editorial feedback I receive.
Knowing how to deal with feedback is a skill I would not want to risk losing. It makes me a better writer, which is what this is all about.
Photo by Collab Media on Unsplash
No comments:
Post a Comment