“What kind of science fiction dystopia are we living in?” grumbled Max. “You’re already using a robot to write your articles for you. And now you want Robot Number Two to check Robot Number One’s work?”
“Well, it’s not exactly like that,” said Eli. He wished he had asked a different colleague for advice. But everything at the CrazyBusy Agency was crazier and busier than usual and no one else was available.
Normally, Eli used AI only to brainstorm ideas and write initial drafts. But this week, he had so many deadlines that he was using AI to write everything. Now he needed to check all of that AI-generated content before he could send it back to the CrazyBusy Agency’s clients. Just this once, he thought he might try using a different AI tool to check the first AI’s work.
Eli should have guessed that Max wouldn’t like the idea. Although they were both in their early thirties, Max was already a grumpy old man. He hated AI and never used it. He would rather stay up all night working at home than use “one of those blasted robots” to save himself a little time.
“Look, I understand it’s nice to eat a meal somewhere other than at your desk,” Max told Eli. “But have any of your favorite AI robots ever churned out an article that didn’t need a bunch of changes before you sent it back to the client?”
Eli thought for a moment. “Sometimes I get pretty good results with the AI tools I use. But yeah, there’s always something that needs to be fixed. And sometimes the AI really misses the mark.”
Max made a sweeping gesture with both arms. “So if Robot Number One’s article turns out to be a huge mess, what do you think will happen when Robot Number Two checks it? Is that second robot capable of working miracles? Or will it just make an even bigger mess? And then at some point, you’ll need a real person to clean it all up.”
Although Eli didn’t share Max’s negative views on AI, he realized his colleague had made a valid point. AI tools could be useful for many tasks, but using them to do both writing and editing probably wasn’t a good idea.
Maybe you’ve been in a situation like Eli’s where you didn’t have time to edit AI-generated writing yourself. Using your favorite AI tool (or a different one) to do the job for you might seem like an easy fix. But this solution comes with many drawbacks. In this article, we’ll take a look at some of them.

Writing mistakes
Most AI tools do a decent when it comes to spelling and grammar. They don’t have as many problems with comma splices and run-on sentences as they did a few years ago. However, keep in mind that AI systems generate language based on the data they were trained on. If that data contains misspellings or grammatical mistakes, then these errors might appear in the writing the AI tool generates for you. And if you didn’t type something correctly when you wrote your initial prompts, those typos may show up as well.
Occasionally, an AI tool might use words like “lite”, “thru”, or “tonite” which are only “correct” in very casual contexts such as social media posts, advertisements, and text messages. Similarly, AI-generated writing sometimes contains short, choppy, incomplete sentences. Like this. And this. Again, this is okay for casual writing. You might even find a few of them in this article! But they’re not appropriate for professional or academic contexts. So if you’re using an AI tool to check AI-generated work, it might let these less-than-professional writing features slide.
Generic, repetitive language
In today’s fast-paced world
Not only _____, but also ______
Whether it’s _____ or ______
AI tools tend to overuse certain phrases and expressions like the ones listed above. These expressions aren’t wrong, per se. But when they’re used over and over again, it screams “this was written by AI”.
If, like Eli, you’re creating multiple articles about similar topics for the same client, keep an eye on the opening paragraphs. This is where AI tools really repeat themselves. You might end up with five articles that start with almost the same paragraph! Sure, an AI tool is capable of fixing the problem, but only if you spot it first and give the AI further instructions.
Hallucinations
As we detailed in a previous TextRanch blog article, AI systems sometimes invent facts, statistics, quotes, research findings, medical data, court cases, and more. They can also distort or jumble factual information so that the end result is misleading. When this happens, it’s called an AI hallucination.
If you’re using AI to correct AI writing, you are increasing the likelihood that your finished piece will contain hallucinations. Without human oversight, no one will be there to catch and correct these hallucinations before they reach your audience.

4. Bias
Just as AI systems can generate hallucinations, they can also generate biased information. (Our TextRanch blog has a whole article about that, too!) Will the AI tool that’s checking a piece of writing recognize that some of the information in it is biased? Will it make the appropriate corrections? Probably not.
5. Missing information
The quality of AI-generated writing often comes down to how well a human enters prompts into the system. Specific, detailed prompts that include high quality source material will yield better results than something simple like “Write me an article about ballroom dancing.”

Unfortunately, the details you include in your prompts don’t always make their way into an AI-generated text. You might have to enter another prompt, instructing the AI tool to include that missing information. Then later, if you use AI to check your AI-generated text, you might find that something else has been left out. It could be a key sentence, a graph, an anecdote, a quote from an expert… You’ll never know what’s missing unless you check it yourself.
6. The tone doesn’t match the audience
One of the most common problems with AI writing is when there’s a mismatch between the tone of the text and the audience who will be reading it. For example, you want something formal and serious aimed at experts in a particular field, but the AI tool produces a chatty, casual article filled with information that your target audience already knows. Or on the flip side, your goal might be to sell a product or service, but the AI gives you a dense informational piece with too much technical content for everyday customers.
Can an AI tool correct these kinds of problems and fix the tone? Sometimes, yes. But like the other common issues we’ve described in this article, the AI can only fix it if a real person finds the problem first and then tells it what to do.
The AI tool makes the wrong changes
Let’s say you’ve used AI to write a blog article. You glance and it and everything looks pretty good. So you run it through a second AI tool just to make sure. But then you discover that the second AI tool has changed things that were perfectly fine in the original version. Even worse, the changes made the article sound uneven and awkward.
This is what happened to one of Eli’s articles. After his discussion with Max, he decided to do a test run. He took an AI-generated article about basketball shoes and entered it into another AI system. He gave it a simple prompt to check for mistakes and fix any awkward phrasing.
“The second AI fixed things that didn’t need to be fixed,” Eli reported. “It rephrased some sentences in a way that didn’t sound right for our client’s brand. The overall tone ended up being too slick and sales-y. Yes, the purpose of that piece was to sell basketball shoes. But the second AI went a little too far in that direction. We were aiming for a friendly tone, not something that sounds like a pushy salesperson.”
Eli decided not to use AI to check any of the other articles. “The CrazyBusy Agency’s clients deserve better than that,” said Eli. “I’m not giving up on AI entirely. But this made me more aware of its limits. There are some jobs that only a human can do.”

How TextRanch can help
Whether you’re a freelancer or the head of your own “crazy busy” agency, it can be exhausting to check every piece of AI-generated writing. Partnering with a professional editing service like TextRanch can lighten your workload, freeing you and your team to concentrate on other tasks. As Max might put it, you’ll have time to enjoy a meal somewhere other than at your desk!
Our TextRanch editors can refine your AI-generated writing so it reads more smoothly and naturally. If the tone is “off”, we can adjust that as well. Best of all, we are open for business 24 hours a day, seven days a week, so we’re available whenever you need us. We work with individual customers as well as with businesses.
What if you’re like Max and you never use AI? TextRanch provides proofreading and editing services to fit your needs as well. Indeed, we’ve been in business since 2016. We’ve worked on business emails, cover letters, resumes/CVs, articles, academic papers, legal documents, full-length novels, and much more. Most of our customers are not native English speakers, so we can help you if you’re struggling with language-related issues.
To get started, just click the blue box below.
Ask an Editor: A New TextRanch Feature!
Proofreading vs. Editing: What’s the Difference?
Human Editors: Why Do You Need One?
How can I improve my understanding of native-English speakers? Part 1