You’ve probably done it too. You paste something into an email, spot those em dashes, and quickly swap them out so nobody thinks a bot wrote it. I’ve caught myself doing the exact same thing.
Then I read a post from a LinkedIn creator who writes for a massive audience, and it stopped me cold. The author’s argument is simple: you’re fixing the wrong thing. Deleting em dashes won’t save you, because there was never anything to hide in the first place. I love a good myth-buster, so let me walk you through the misconceptions this expert takes apart.
The myths people believe about AI detection
Most people assume AI detectors work, that a stray em dash exposes them, and that the goal is to make writing look “human.” The original poster shows why almost all of that is wrong.
- Myth: AI detectors actually catch AI. They’re basically coin flips. The author points out that OpenAI shut down its own detection tool because it couldn’t reliably detect AI text.
- Myth: Detection tools are trustworthy science. The creator notes that companies pull in serious money selling this dream, and calls the whole promise a scam.
- Myth: Short text still gets flagged accurately. Feed a detector a few sentences and it has nothing to work with. It’s pure guesswork.
- Myth: The “AI look” is permanent. Prompt the model for a specific voice or style and that robotic tone disappears instantly.
- Myth: People get caught for using AI words. Nobody got flagged for writing “delve.” As the expert puts it, they got flagged for zero effort.
That last point is the whole game. It’s not the vocabulary. It’s the effort.
The real problem: workslop
Here’s the truth this industry pro wants readers to act on. People are busy hiding their AI, and very few realize there was nothing to hide. The real enemy has a name: workslop.
The author defines it as work that looks finished but quietly dumps the actual effort back onto you. Think of the polished-looking report that falls apart the second someone has to use it. The tidy summary that misses the one thing that mattered. That’s what people actually notice, and that’s what damages your reputation. Not punctuation.
You’re a better detector than the expensive detection bots. Humans spot missing effort instantly, even when a machine can’t.
I think this reframe is worth its weight in gold. We’ve been so worried about looking human that we forgot to actually do the work.
How to put the effort back in
You don’t need a course to apply the lesson from this post. Here’s how I’d act on the creator’s insight starting today:
- Stop deleting em dashes. That energy is wasted. Nobody’s running your emails through a scanner.
- Read your AI output like a skeptical reader. Ask if it actually solves the problem or just looks like it does.
- Add the missing 20 percent. The context, the judgment call, the specific example only you would know. That’s the human part.
- Prompt for substance, not disguise. Instead of “make this sound human,” ask the model to pressure-test the idea or fill the gaps.
- Own the final pass. If you’d be embarrassed to defend it in a meeting, it isn’t done.
Why does this matter beyond your inbox? As more teams lean on AI, the flood of polished-but-empty work is only going to grow. The people who stand out won’t be the ones who hide the tool best. They’ll be the ones who use it to ship work that’s genuinely better.
The person who shared this said it perfectly: forget hiding the AI, and put the effort back in. That’s a filter I want to keep using.
Want the full breakdown, including the reasoning behind why detection fails? Check out the original LinkedIn post for all the details.