Skip to main content

Command Palette

Search for a command to run...

I Was Skeptical of AI. Then It Saved Me Hours of Writing.

Updated
โ€ข9 min readโ€ขView as Markdown
I Was Skeptical of AI. Then It Saved Me Hours of Writing.
A

Ayu is a tech blogger, open source contributor & maintainer, and tech community enthusiast based in the Netherlands. She loves photography and iced coffee!

Hi friends ๐Ÿ‘‹,

I've been writing tech blog posts for a few years now. If you don't know me yet, English is my second language. My mother tongue is Indonesian, and I'm also surrounded by Dutch every day since I live in the Netherlands.

Being a polyglot means my brain works in a funny way. Some days, I can speak or write in English or Dutch without any effort. Other days, I have to think in Indonesian first, then translate. Sometimes my English is stronger, sometimes my Dutch is. But there are also times when nothing comes to my head except Indonesian.๐Ÿ˜…

That said, writing with all that going on in my head has never been easy for me.

I've been wanting to write this post for a while now because I used to feel skeptical about AI, but one experience changed everything.

My Old Writing Process

Before I opened up to AI, here's what writing a blog post looked like for me.

  • Jot down ideas in bullet points
  • Struggling to write the first sentence
  • Struggle even more with the first paragraph
  • Find the groove and get into a flow
  • Get stuck again
  • Step away for a bit and do some house chores (or play mini games like Minesweeper ๐Ÿ˜Ž)
  • Come back and repeat the cycle until the post is magically done
  • Run it through Grammarly to fix the grammar and typos
  • Run it through the Hemingway app to check if it reads easily
  • Throw sentences into a translator app when I forget how to say something in English, or to double-check that it makes sense

And that's not counting research, like googling things or reading other articles to check my facts. A task that takes most people 1 to 2 hours took me more than 4 hours, just for 800 to 1000 words. That doesn't include the time to make a cover image.

Why I Was Skeptical

For a long time, I didn't trust AI. I was afraid people would think my writing "sounds like AI" and stop trusting me. I was afraid I'd accidentally fed it something I shouldn't. I was afraid it could act on my behalf or access my computer in ways I didn't want. The list goes on.

On top of that, I couldn't afford even the cheapest AI subscription to try the better version.

So while people around me played with AI, and some even became experts at it, I stayed on the sidelines.

watching gif

Wait, Was I Already Sounding Like AI?

Then something made me pause. People started to say that certain phrases indicate AI writing, such as "let's dive in" or "that being said," as well as using too many em dashes or too many bullet points.

But here's the thing. I've used bullet points and em dashes long before AI became a topic (I even mentioned it in some of my blog posts). As an organized person, I love them! I say "let's dive into" and "that being said" often, too. So was I sounding like AI this whole time? Or did AI learn to write the way people like me already do?

Same with emojis. I'm an expressive person in real life, so I use them often when I write. People used to say AI can't read and do emojis. Now, try asking AI something, and it might throw a few emojis right back at you. So what does that make me, then? ๐Ÿ˜›

The Day I Gave In

One day, I had a blog post to write and publish fast, with no time for my usual 4-hour process. I gave in and used Gemini. I did what I always do. I jotted my ideas in bullet points, then asked it to draft the post.

In less than 5 minutes โ€” although with some robotic tone โ€” I had a draft! A task that normally took me 4 hours (sometimes days if I got stuck finding the right words, even with my ideas right in front of me) took me minutes. With editing and making the cover image (yes, it still needed editing), the whole thing took me less than an hour and a half to finish and publish.

I remember thinking, "Wow. I still have plenty of time to do other things I need to do today. I'm so productive!" If you're a mom โ€” or a dad โ€” with little kids, you'd understand that even having 5 minutes of extra time to get something done before your kids need your attention is priceless!

Your Ideas Are Still Yours

Here's something that hasn't changed for me, before or after AI. I despise plagiarism. Copying someone else's work, changing a few words, and passing it off as your own without giving credit has always bothered me. People put real time and effort into their work. When someone heartlessly steals that labor, it is not okay โ€” AI or not.

However, using AI to draft your own ideas is a different thing entirely. Many people can write about the same topic, and it can still be original because everyone communicates in their own way. I've always told people who are just starting to write: write what you learn, write the solutions you find, write about what interests you. Other people might cover the same topic too, and that's fine.

Where I Draw the Line

That brings me to another point. Does using AI mean the ideas aren't yours anymore? I don't think so. AI can't read your mind. Even if you give it one prompt like "write me an article about X," it still doesn't know your direction. You have to guide it, step by step, with your own words and ideas.

So where's the line? For me, it's this: "Assist me in drafting these ideas" is not the same as "write me this topic," then closing your eyes and hitting publish. If it's your idea, if you know where the information came from, if you actually read and edit what AI gives you, you can confidently stand behind your words when someone asks about them. That's what matters. I see people cross this line many times in my daily life as an open source docs maintainer.

Since I started using AI as an assistant in writing, I've learned to notice its patterns. I can immediately spot when someone is using it irresponsibly. Like, they'll submit a PR and won't even take a second to check if the markdown format is broken, if the style guide is ignored, or if the content is even correct. Itโ€™s just so easy to forget that behind every single PR โ€” whether it's docs or code โ€” thereโ€™s an actual human maintainer who has to spend real time and energy reviewing it.

I don't oppose the use of AI. But I can't emphasize enough how important it is to be responsible. That means you need to check what it threw at you. Check the format, check the style, check the content, and double-check the facts before you hit that submit or publish button. Because being irresponsible doesn't just cause someone else to waste their valuable time, it also makes people lose trust in you.

Being Transparent

So here's my honest disclosure. I used AI to draft this very post, based on my own ideas and experience, which I gave it in bullet points, in no particular order.

That's how I usually come up with ideas. Messy, out of order, sometimes hard to connect. Like, I know something is important, but I can't fit it in, so I cut it. Not because it doesn't matter, but because I struggle to make it flow.

AI helps me keep that idea and makes the whole article make sense. Even so, the work never stops there. Like right now, an hour after the draft was finalized, I'm still sitting here editing the content and fixing the grammar (even AI can't escape Grammarly flags). Relying solely on AI isn't an option.

Reality Check

Out of curiosity, I omitted the "Being Transparent" section and tossed this whole post to Claude in the incognito chat (so it doesn't know that I'm the same person). I gave it this prompt: "As an AI, what is your thought about this post? Is it written by AI?"

Claude's verdict? It told me the voice felt too real and personal to be AI, pointing out that the slightly messy structure and 'small imperfections' actually read as human.

Intrigued, I tossed the text to Gemini as well. Here's what it said:

The Verdict: Human Heart, (Admittedly) AI-Assisted

It reads like a genuine personal essay written by a human who likely used AI exactly how they described: to break through the language-barrier friction, which they then heavily edited to inject their own voice, specific life context, and personality.

Seeing an AI confirm that I don't sound like an AI is kind of hilarious. But it shows that when you use your own ideas, the human part still cuts through the noise.

Final Words

I might be late to the party. But I'm willing to open up and give AI more chances, especially on days when I'm stuck or have too many ideas and too little time to get them all out. ๐Ÿ™‚

That's my side of the story, but I'm curious about yours โ€” especially if English isn't your native language. Do you go through the same kind of struggle when putting posts together? And would you (or do you) use AI to help you write? I'd love to hear how you handle it in the comments!