Artificial intelligence is dominating conversations among writers and content creators. I tried to stay out of it—until it became impossible to ignore.

At its best, AI can be a useful tool, particularly in nonfiction. But the real issue isn’t whether AI can assist writers—it’s that it’s now being used to judge them. And that’s where things become ethically and professionally dangerous.

An ethical writer verifies sources before publishing. That’s the baseline. But a rushed or disengaged writer might lean too heavily on AI. The irony is that even careful, ethical writers can now be penalized—not for cheating, but because flawed AI detection tools say they did.

I experienced this firsthand in a recent writing contest. The judge ran all submissions through an AI detection program and rejected most of them. Out of more than 900 entries, only a little over 100 made the cut. That number alone raised questions. Are we really expected to believe that nearly 800 writers relied heavily on AI?

Curious—and concerned—I tested the contest’s detection tool myself. I submitted a chapter I had written long before AI was even part of public conversation. One program claimed it was 38% AI-generated. Another put it at 89%. The same chapter. As the author, I know it was entirely my own work.

That’s not just inconsistent—it’s alarming.

Writers seeking agents, publication, or recognition now face a troubling possibility: that their originality could be judged by unreliable software. For industry professionals, these tools may seem like efficient shortcuts. But they’re far from dependable, and when they’re wrong, the consequences fall on the writer.

The most unsettling moment came after one of these programs flagged my work as AI-generated—then offered to rewrite it “to sound more human.” That contradiction says everything. A tool that misidentifies human writing and then claims it can humanize it exposes its own limitations.

Concerns about AI aren’t new. Early development relied on massive datasets, many of which included copyrighted works used without clear permission. That ethical issue remains unresolved. And yet, despite these concerns, AI has been introduced rapidly, often without transparency or fully tested safeguards.

We’ve lived with forms of AI for years—automated chat systems, customer service bots, phone menus. They were imperfect but contained. Now, their reach is expanding, sometimes beyond what current systems can reliably support.

I saw this up close in a situation that had nothing to do with writing. When our washer broke, we ordered a new one from a nearby store. What should have been a routine delivery turned into a week-long ordeal. AI systems misrouted calls, transferred me to a warehouse in another state, and ultimately led to the cancellation of my order—because the system decided Louisiana and Oklahoma were interchangeable.

It took persistent effort—and finally, a human customer service manager—to fix the issue. Even then, another error slipped through: an incorrect zip code delayed delivery again. Each time, the solution came not from automation, but from human intervention.

This is the core problem.

AI can reduce workload. It can streamline processes. But it cannot replace human judgment, especially when nuance, accountability, and real-world context are involved.
Writers aren’t just facing competition from AI—they’re facing evaluation by it. And when the tools doing the evaluating are inconsistent or inaccurate, the risk isn’t theoretical. It’s personal.

Careers can be damaged. Opportunities can be lost. All because a machine made a decision it wasn’t qualified to make.

AI is not the enemy. But neither is it infallible. It is a tool—nothing more. And like any tool, it requires oversight, skepticism, and responsibility from the people who use it and present it.

Until we draw that line clearly, we risk allowing flawed systems to make decisions that should always belong to humans.