When AI-Generated Content and Human Sensibilities Intertwine 

Should I feel ashamed that I chose my company tagline based on a ChatGPT suggestion?

When the whole AI-generated content thing reared its ugly head, I resisted. I thought, surely this will result in crappy, soul-less content and is just morally wrong.

But I was tempted. While I wasn’t ready to sell my soul to AI, I was prepared to take a cautious nibble of the forbidden fruit. It just so happened that one of my first experiences was surprisingly positive. I needed help with my tagline and asked ChatGPT to create several versions of it, incorporating the concepts of weaving (tying to my last name), strategy and creativity. 

I won’t bore you with the long list of marginal-to-bad suggestions from my AI brainstorming partner, but by combining some of the concepts and verbiage, I was able to come up with a tagline I like and that I feel highlights my strengths―along with the weaving concept:

“Where Strategy and Creativity Intertwine”

I admit, it was helpful to have the “collaboration” of ChatGPT.

A Starting Point

The typical advice to content writers is that AI-generated content should be a “starting point” for writing copy. This seems like a reasonable suggestion. 

It’s kind of like using a template, I thought, which is another kind of starting point. Here’s the caveat: As someone who has created scores of templates for use across organizations in many different hands, I am always stunned by how badly a template can be screwed up. 

For example, a nicely designed flyer template in the hands of someone with no design skills can still end up looking amateurish, at best. Without the talent and experience to properly use the starting point, the results are likely to be less than professional. 

Extending this metaphor, the same should apply to writing with AI – especially if the “writer” is not a writer. No matter how well-engineered the prompt (the starting point) may be, the quality of the final product depends on the human’s ability to judge the quality of the output.

Early Observations

Here are a few takeaways based on my personal experience so far on using AI for content writing assistance:

• As a solopreneur, working from my home office in the company of only non-human, non-verbal Golden Retrievers, I appreciate having access to “someone” to bounce ideas off of―even if they too are non-human.
• ChatGPT is great for research, and for that, a huge time saver. For me, at this point in time, research may be its most useful application. Granted, there is a serious need to fact check. I’ve gotten some amusing―and alarming―results using ChatGPT to research topics that I know well. It’s a good object lesson in whether to trust the accuracy of AI-generated content.
• There is very little AI-generated content that I would feel comfortable using without significant modification. It would feel lazy, like cheating. Besides, I want to write original content in my own unique human voice. 

Reserving Judgement

I admit, I’ve never been an early adopter of technology, but nor have I been in the late majority or a laggard. I’m somewhere in the middle, at the top of the bell curve. For now, I’m reserving judgement as we collectively work toward finding the right place for AI in copywriting, content writing and the creative process, in general.

It will be interesting to see how AI-generated content plays out. Tom Mangan, a freelance writer and frequent contributor to LinkedIn discussions on AI, posted the astute observation that AI models were “trained on human-generated content, not bot-generated content.” This is likely to change over time, he said, as GPT-generated results are pasted into articles and published, eventually corrupting the training of generative AI. 

Is there a risk, I wonder, of ending up with only bot-generated content, many generations from the original human voice?

While it’s tempting to rely more and more on AI-generated content, as Tom points out, “It’s akin to switching to an all-coffee-and-cookies diet―it feels great in the short term but kills you in the long run,” he said.

I wonder, just how many iterations will it take for content to lose its humanity?