Ugh. Have to talk about it. I first tried AI after I’d written that last of my series. I NEVER used it to generate any aspect of my stories. Sad I have to declare that, but as usual, people who don’t abuse things sometimes suffer the consequences of those who do. Meaning, now any use of automated intelligence (AI) can make an author the target of the AI witch hunt. Just know this: If a story or article is by Beth Maddaleni, I wrote it–not AI.
What I’ve tried/What I’ve found
Within the past year, I’ve tried Claude, Marlowe (by Authors A.I.), ChatGPT, and AuthorCraft. Marlowe analyzes a manuscript, and offers insights and graphs about it. It also lists cuss words, typos (which aren’t always typos), and other terms it categorizes for one reason or another. Claude and ChatGPT pulled and listed information from my manuscripts that I directed them to pull. AuthorCraft offered me developmental insights.
I’ve since deleted AuthorCraft and stopped using Claude. I haven’t used Marlowe in a while but would consider using it again for manuscript analysis. ChatGPT is my go-to as a search-engine alternative. I like that I can click on sources that look legit and then decide for myself which to use, if any at all.
I once generated a lovely video using AI images I directed ChatGPT to make but have since deleted it from TikTok along with another video that was a mix of human and AI art. I deleted them out of fear after seeing a person’s negative comment about AI. I worried Would I insult real artists? Would I ruin my reputation as an author? Whether my decision was right or wrong is probably in the eye of the beholder, or laws. For now, I doubt TikTok would have allowed me to post those images if doing so had been illegal. However, I don’t know enough about AI-generated images to say whether posting them is at least morally right or wrong. In doubt, I nixed them.
What I discovered
No AI is perfect.
- I’ve used it to pull info from my manuscripts: finding typos, inconsistencies or having it compile story elements into a story bible. AI did not execute those tasks perfectly. Sometimes it sees an error that’s not there or reads a made-up term as an error. One time, it categorized the wrong person as my MC’s mother.
- A responsible author (hello, that’s me!) knows s/he can’t solely depend on AI for editing tasks, and asks the bot to not write any content, just answer questions or list whatever.
- When I ask it if I’ve used a word correctly, I don’t trust it. More so, I trust Merriam-Webster and primary, living sources, and same for my research. If I do ask AI a research-type question, I click on its sources before considering the information valid.
- AI can’t outshine a humans!!! I won’t stop seeking human insights, editing, and artwork.
In truth, at least some authors should test AI, as I have. With no experimentation, how does anyone in the writing and publishing industry have a right to say whether it’s good or bad? AI can be used as a starting point. As mentioned above, you can ask it a question but can’t trust all the information it offers. You must go to the sources it provides, then weed them out for the most reliable, such as primary or official sources. More important, seek human experts when possible.
*My comments about the above AI products are opinion based on my experience with them the past year. I’m not endorsing any, nor am I trying to influence your decision to use them or not.
I remain cautious but realistic about AI. My gut usually tells me what’s right and wrong, and it hasn’t failed me yet. So I think it’s okay to use AI responsibly, in a way that doesn’t generate content. My writing will always be authentically mine.