What AI Can’t Write For You

By Seanice Kacungira

I write all my articles myself.

It’s a form of catharsis and one of my favourite things to do when I feel like my thoughts are all jumbled up with no real clear line through them. Writing garbage at first, emptying all my ideas out on the page, feels a lot more satisfying than asking an AI to come up with something neat and packaged. There is something very important that happens to your brain in the process of writing. A refinement, a sharpening of your skill set in a way that can be frustrating and glorious all at once.

You sit back to read what you wrote and you gasp in horror. Unco-ordinated sentences stare back at you. Incoherent thought patterns with no real thematic engagement in them. And then, slowly, you find the thread.

A friend sent me an article to read this week. It was obvious within seconds that it was a few of her thoughts written and deciphered by AI. In ten seconds flat I spotted the first discrepancy, a failure to support her assertions with any real data. What she purported to be true was so easily debunked it was almost laughable. A rookie mistake. And yet I know this friend to be thorough in most aspects of her life. So what happened?

Was she simply seduced by the ease with which the words flowed and seemed to make sense, so much so that she didn’t stop to engage her brain and ask: do I believe this? Is this assertion beyond reproach?

Here’s what I’ve come to understand: even if you use AI to write, research, or strategise, if you know your subject matter well, you will see the gaps almost immediately. You will be able to question. The problem isn’t the tool. The problem is arriving at the tool before you’ve done the thinking.

I remember when I was a young scholar, my favourite literature teacher told me something so similar to what we tell kids about the internet today that it will make you chuckle. He said to me: “Don’t believe everything you read.”

Ever since that day, I started developing a voice inside me that said exactly that every time I picked up a self-help or autobiographical book. I’d find myself asking: do I believe this author’s view of the world? Does it resonate with my lived experience? Is this how I would like to show up in the world? Do these words align with my value system?

You might wonder how one does that with autobiographical books, a person’s account of their own life. Well, when reading those, I would ask myself: what part of this person’s life gripped me, and why? What could it teach me about will, perseverance, character? Is there a chance this person is embellishing, giving themselves more credit than is due? Is there a consistent pattern to who they are, or not? What role did plain old luck play?

In doing so, I was never just a passive reader. I was an active co-writer and creator, a real live thought partner. And my reading was never the same again.

That is exactly what applied thinking with AI looks like.

Don’t believe everything it tells you. Develop your own writing style, your own thought patterns, understand your own cadence. Do the hard work first. Because when you arrive at AI having already wrestled with your ideas, something different happens. You stop consuming its output and start interrogating it. You become the editor, not the audience.

And this, I think, is the most underexplored possibility of this whole moment: AI doesn’t just make our lives easier. Used well, it can open the doorway to deeper work, a deeper interrogation of who we are, what value we bring to the world, and who we eventually want to become. But only if you bring yourself to the conversation first.

So write first. Then ask AI to criticise you. And maybe, just maybe, you’ll learn a thing or two and sharpen what you already know to be true.

(Yes – This article was written all by myself!)