On Using Generative AI

On Using Generative AI

A few weeks ago I was contacted by a magazine whose target readership is independent publishers much like Central Avenue. They were writing an article on generative AI and its applications in publishing. As someone who uses it for work, I shared my thoughts with them, and the article should come out later this summer.

This is a really tough one, because I feel as though no matter what I say, I might be raked over the coals. Nevertheless, I wanted to share how I’ve been using it at this small press.

Firstly, I want to highlight that I am far from being an expert or having a deep understanding of generative AI. However, I have a keen interest in it, and enjoy learning from people more knowledgeable than I am. To begin, if you’re wondering what generative AI is, take a look at the difference between predictive and generative AI. If you’d like a slightly deeper dive, I suggest this podcast by Cal Newport, a modern thinker whom I admire. 

Now that you’re as much of an expert as I am, I’d like us to keep a key word in mind: “yet”. I was at a publishing conference a few months ago and the sentiment in the room was that generative AI is not something that’s capable of major disruption, taking away jobs, or posing a danger to humanity. I mostly agree with this view, but with one caveat: if you add “yet” to any of those concerns, you’ll have a good idea of how I feel.

But ChatGPT is here along with its many iterations and it’s being widely used in a variety of uses, from the fascinating to the nefarious, and to ignore it is wrong on many levels. For me, I use it, and it greatly benefits my workweek. 

Here are my main uses:

  • Generating and rewording marketing copy: To sell books to the trade, I have to come up with so many sell lines for so many different advertising and promotions for so many books and I’m not overly creative or equipped with a marketing brain. I don’t have an in-house marketing guru and the copy that ChatGPT has come up with sounds pretty darn good to me. If it wasn’t hard enough for me to come up with the original copy, now  a particular campaign dictates I can only use a 50- or 100- word blurb. Using AI helps me edit that copy into something just as engaging – in a few seconds.
  • Proofreading blog posts & newsletters: I’d like to think that my writing is flawless but it isn’t. And yes, I use AI to help proofread these newsletters and blog posts. When I do use it, I go back over the output to make sure it still sounds like me (grammar mistakes and all!).
  • Designing presentations: AI won’t build my powerpoint slides, but it does come up with some great ideas for slides that I might not have considered.

What I won’t use it for:

  • Confidential or copyrighted material: I have not read the general T&C of  ChatGPT, but I can assume that anything I feed into it becomes the property of the AI. So I won’t put in anything that may unintentionally grant the license for that material to whichever company that owns whatever AI.
  • Editing a book: I will not use it to copy edit or proofread any book I publish. The reality is that humans are much better at this task. And perhaps that capacity is coming, but I don’t see it (yet).
  • Anything an artist can do: Much like I tend to use the checkout lane with a human cashier, I won’t use it to generate verse, prose, or visual art since I prefer to have a human generate it for me.

So to sum up, I basically use it for admin and marketing. And yes, I know those are jobs that humans do, but the reality is that I simply can’t afford to hire a whole human to do those small tasks. By using AI, I save money, which enables me to acquire more poets and authors, as well as hire visual artists to illustrate our covers.

I’m very curious to see how this all plays out in the industry, because I believe that it will take on a much larger role. And if people claim they aren’t using it – well, I’ll simply add a “yet” to that statement.