Software Doesn't Replace Judgment. It Never Did.

On Substack's new checker, a flagged comma, and what actually changes when you use the tool well

Eduardo Ramos for Unsplash+

Substack rolled out something new this past week. Every post now has an AI Checker button, so readers can click it to see an estimate of how much of what they’re reading was written by a human versus AI. It’s an interesting idea. I just wonder why it became necessary.

My own writing has been flagged for AI before. I can only assume it’s because I use proper punctuation, because it’s certainly not my sentence and paragraph structure. So I have to ask, has AI become enough of an issue on the platform that readers need a scanner to sort it out? Is there still an em dash witch hunt going on here? And is Substack ironically using AI to test for AI? Because that would be hilarious.

In my world, AI stopped being a debate a while ago. I use it every day for technical support, research, proofreading, meeting notes, and more. Writers have been using AI for years through Grammarly, spellcheck, predictive text, and nobody thought twice about it. The tools have simply become better and the names have changed.

I’m absolutely against using AI to write an entire novel and passing it off as your own work. That’s where the romance author featured in the New York Times in February crossed a line for me. She used Claude to generate hundreds of novels and repeatedly skipped disclosing AI use on KDP because “AI books don’t sell.” The dishonesty and flagrant KDP TOS violation is maddening.

We’re hiring at Plotsy, and I’ve been interviewing candidates for the past few weeks. Before every interview ends, I ask the same two questions: How do you feel about AI? And how do you use it personally and professionally?

The answers have been overwhelmingly positive. Maybe that’s because it’s an interview. Maybe it’s the kind of work we’re hiring for. Either way, I’ve walked away from more than one conversation with a new idea I hadn’t considered before.

Professionally, everyone I know uses AI to some degree - research, planning, organizing ideas, building agents, coding, ad campaigns, custom GPTs. It’s become just another tool we use to get work done.

I use AI to help build editorial calendars around the projects I’m managing that week. It helps me spot patterns in meeting notes and develop strategies before small problems become bigger ones. It saves me time, but it doesn’t make decisions for me. It still doesn’t sound like me without a lot of editing, which is kind of the point.

As a tool, AI doesn’t work by itself. It depends on someone to ask good questions, recognize bad answers, correct mistakes, apply unique experience and voice, and decide what gets published. That’s no different than opening Photoshop, Microsoft Word, or Excel. The software requires human judgement and can’t replace the person using it.

I ran this through the checker before publishing. 100% human. Make of that what you will.

Lynn McGinnis writes about publishing, author visibility, discoverability, and the future of building a writing career. As Founder and Chief Publishing Officer at Plotsy, Inc., and with more than thirty years of marketing experience at Lynn McGinnis, LLC, she brings practical insight to the changing world of books, technology, and author platforms. Her work helps authors think differently about visibility, publishing, and building careers that last. She writes from her little lake house in rural Pennsylvania, where she lives with her husband, son, and two fluffy companions, Kaiju and Bella.