Using A.I. in Books

A.I. can create beautiful images.

Except the woman’s hand does not look quite right.

For my book, Becoming the Enemy, I thought it would be fun to illustrate it with A.I. images. (Becoming the Enemy is science fiction and not a romance.) I found it very frustrating because it took many trials for me to find what I could use. I decided to create some examples to show some of the problems.

For the fun of it, I tried the prompt, “Darcy and Elizabeth.”

Elizabeth is shorter than Jane and Lydia. Darcy is tall. Are Jane and Lydia as tall as Darcy? Is Elizabeth standing on a stool?

Here are two pictures, out of four, I got using the prompt, “A regency era woman wearing a gown with a high waist standing in front of a table.” (Yes, I know “Regency” should be capitalized, but I don’t think the computer cares.)

She isn’t standing and her dress is not high waisted.

She is standing, but not in front of the table. Her gown is not high waisted.

None of the gowns were high waisted. Perhaps A.I. doesn’t understand “high waisted.” So I googled “empire waist” and got, “An empire waist is a style of clothing where the bodice is fitted and ends just below the bust.”

I used the prompt, “A regency era woman wearing a gown with an empire waist walking briskly with a dashing gentleman in a garden.”

This does not look like a couple walking briskly. Also, I don’t think Regency women would wear their hair down like this.

I was never able to get a high waisted gown.

Here is an image I used in Becoming the Enemy despite the rather weird left hand the woman has. Daisy is a minor character, but I think I would have mentioned it if she had such an unusual hand.

I was hoping no one would notice the hand.

My prompt was, “Two men and two women playing cards in the Regency era with a militia lieutenant observing.”

Where are the observer and the second woman?

In the following image, I think the man’s right hand appears twice and I’m not certain what is going on with the woman’s left hand.

I’ve been asking for four images and only using one. For my last example, I’m giving all four. My prompt was, “A carriage with matched bay horses is driving through Regency London.”

This doesn’t look like it will work.

I’m not certain what is attached to what.

It is customary with one horse to put it in the center. Also, having one horse simplifies the problem of finding those that match.

Is there a carriage? What is the horse on the left attached to?

Perhaps A.I. will take over many jobs, but it still needs some work.

I hope you will try reading Becoming the Enemy where I was able to get pictures I could use.

Leave a Reply