Generated content gives me the heebie jeebies: my take on the AI conundrum

I studied languages and literature because I wanted to be a writer. There’s nothing quite as beautiful as a perfectly composed pitch, poem, ad, or landing page. So when generative AI showed up in content, I didn't just distrust it. It irked me immensely, and it stayed that way for about three years. Come with me on this journey through time and space, and see where the issues popped up, and how I built a bridge to overcome my avoidance of change.

From a masterclass in lazy prompting…

In the early days, purely generated content was easy to spot. Generic words, generic phrases, and not a single thing of value in the entire piece. No insight, nothing you didn't already know, nothing that made you want to keep reading. It was obvious, boring, repetitive, and full of emptiness.

Then I realised half of the team I was managing at the time was using AI, or rather, relying on AI to do the writing. Here’s what the process looked like:

• Poorly written prompt

• Brief copied and pasted into ChatGPT

• Copy/pasting the generated content and submitting

And the funny part? I didn’t realise that this was their process for quite some time. It popped up in the empty phrases and obvious intros. And not to mention the “excitement” and “diving” that went on, and on, and on… Yes - I noticed a pattern. I then thought to check what was happening, and that's when I noticed the laziness that had taken hold. And that was when I put a policy in place. Writers were welcome to use AI for inspiration, but not to rely on it and hand in whatever it produced. I flagged generated content constantly and sent it back, again and again.

My reasoning is and will always be that high-quality content written by humans, for humans, content that actually matches the user’s needs (or intent), will rank, perform well, and ultimately result in conversions. A poorly written prompt, coupled with a basic brief or template fed to AI, will always spew out the most obvious version of the brief.

I’ve since come to the realisation that my actual problem or annoyance with AI was not the tool itself, but rather how I was introduced to it. One could go as far as to say I got a masterclass in how NOT to use it.

…to figuring it out

After a couple of weeks of begging and pleading, and still refusing to touch AI myself, I started doing weekly creative sessions, with the hope that sparking the creativity would lead to better prompting. This strategy worked quite well, and I started seeing less “exciting” and more substance over time.

But what about the moment I started implementing AI within my own process? I wouldn’t say it was that I just bought into it (while I love learning and challenges, my passion for words and storytelling just wanted to avoid it), but rather that my workload asked for it. And the people who truly know me, would attest to the fact that I am a workhorse who would overload myself with manual tasks that could have been automated.

I then started prodding and testing and figuring it out for myself. First with simple prompts, asking AI to check an edited brief alongside examples and competitors and highlight any gaps. Then, I started using AI to speed up my reporting. This was one of the things I used to spend a lot of time on (and typically would keep me working till 7pm on a Friday). Once I got the hang of it, I tested out a QA prompt, which cut the time that took process in half. 

In the end, I came to see that great prompting and using the tool discerningly while purposefully would streamline things, free up time, and deliver reasonable results. I would still not (to this day) rely on any AI tool to just generate something meaningful without any human oversight or with lazy prompting for that matter.

It’s not about the tool, but rather about how it’s used

Lately I’ve been seeing a lot of ‘lazy’ content - e-very-freaking-where - it’s not even spun, it’s just generated. And, as you’d imagine, there’s a lot of talk within content circles about some companies hiring human writers again after laying them off and trying to rely on AI for content production. On the flip side of the coin, there’s a lot of folks who praise AI. 

I can’t really say that I’m on either side to be honest. I use AI discerningly for tedious tasks, and I keep it well away from the ones that require critical thinking or human connection. There's real value in it. I just don't believe replacing writers with AI entirely is the answer.

My take is that, much like any tool out there, it can add real value, no doubt about that. But, in order to get to that value, you need to know how to use it, and use it well.

My advice to writers

Don't use AI to fill a blank page. That's the fun part of being a writer, after all. If you struggle with structure or flow, use it to hint at those, or to adjust and tweak what you've already written. That will always beat a lazy prompt that produces thin, empty content.

And learn to prompt. Keep developing that skill, because the output will only ever be as good as what you put in.

The “hill I'm still prepared to die on” hasn't moved: the human bit matters. What changed is that I stopped treating AI as the enemy of it, and started treating it as a tool that's very bad at that part and quite good at everything around it.

If your team is figuring out where AI fits (and where it doesn't), that's work I do.