Dirty Business

There are times when I think that I'm going to receive a knock at my door with a couple of police officers wanting to question me. Or worse, that I suddenly show up on a no-fly list.

Honestly, I wouldn't harm a fly.

When I was doing research for Dark Water, I Googled all sorts of sources about the decomposition of a body, the plausibility for a murder, and all sorts of forensic and legal questions, as though I was trying to get away with a crime. The very first time I presented a scenario to ChatGPT, it responded by saying that it could not give me the answer to my query because the scene I described was of a criminal nature.

I had to explain that I was gathering this information for crime fiction, and laid out a blurb of the story before the AI tool understood my motive for these questions, to which it then gave me all sorts of help.

I'm back at it again for The Watcher, and I'm finding that ChatGPT is still showing reluctance in providing me with details. And perhaps my questions and scenarios are too explicit for the tool to want to be too precise.

The dead body in Dark Water is pretty straightforward: a woman is found floating, face down, in the Rideau River. The victim could have merely drowned.

(Without me giving too much away) in The Watcher, the murder victim is more deliberate, more obvious. A chocolatier is found in her ByWard Market shop, arranged in a tableau on the floor, positioned with purpose.

Image: ChatGPT

Because the position of the body is not natural, the killer took great pains to arrange her. And so, I've been asking various AI tools if my hypotheses are correct.

Reading the answers that ChatGPT has provided me, I can almost sense that the bot is seemingly uncomfortable, especially when it says that it can't give me step-by-step verification—not that I'm asking it too: I'm simply addressing possibilities and asking how plausible or realistic those possibilities are. Sometimes, it merely suggests that I don't get into precise details but that I make things somewhat vague.

I argue that the information isn't so that I can put it in the story, but that I simply want to know to help me understand so that I can make the story more believable. And perhaps that sets off alarms.

Depending on the answers I do get from the AI tool, I might modify the narrative. Already, I've made a few changes.

But I can't help but wonder if this research tool might want to report on me, may turn my queries over to authorities.

Will the cops come to my door? Time will tell as I ask more and more questions.

Comments