Donald Firesmith - Author

HUMAN AUTHORS INSULTED BY SOME AIs.

As some of you probably know, I have over 40 years of experience developing large, complex, software-reliant systems. I was also named a distinguished engineer by the Association of Computing Machinery. I do not write this to boast, but to explain why, after retiring from the Software Engineering Institute to devote myself to my first love (authoring speculative fiction novels and short stories), I decided I couldn't completely give up on the field. For that reason, I have devoted the last 5 years to keeping up to speed on AI and robotics. I've even written technical whitepapers on the subject.

To finally get to the point of this email, I have been wondering how long it will take before AI can generate high-quality novels and short stories. My children's book of bedtime stories (The Adventures of Periwinkle Fae) is the result of an experiment designed to give evidence of an answer.

Another thing I have been doing lately is asking the frontier LLMs (i.e., ChatGPT, Gemini, Claude, Mistral, Deep Seek, and Copilot to tell me whether they think my stories were written by a human or generated by an AI (and give their reasoning behind their answer. The results have been interesting and somewhat unexpected.

So I gave my latest novelette (a horror story titled Room 307) to the AIs, and the results were 4 said definitely human, and two said it could be human, AI, or human collaborating with an AI. The two LLMs that were unsure said that the story was too polished and free from errors / weaknesses.

I IMMEDIATELY REALIZED THAT I (AND ALL HUMAN AUTHORS) HAD BEEN INSULTED! THE AIs THOUGHT THEY WROTE BETTER STORIES THAN HUMANS AND THAT PROBLEMS WERE A SIGN OF HUMAN AUTHORSHIP.

The oligarchs and mega-corporations always find ways to shift the blame for economic problems they cause. A little cartoon I came up with this morning.

How to properly frame the narrative.