As we use AI more in business and marketing, there's a real danger that whatever you write or create does the exact opposite of what you intend.
In this post, we'll look at AI slop, what it is, how to recognise it, and why it's bad for your business.
AI Slop - A Definition
AI slop is content, of one form or another, that has been created with no, or very little, human interaction. Someone might provide a title and a short prompt and get an AI to create output with no further human input, and that content gets published or used without human oversight or reasoning.
When used in the marketing world, it's lazy and easily recognised.
In web design, AI-generated pages follow the same layout tropes and colour schemes, with no imagination. In copy, you see "matters" and "quietly" a great deal, and well-structured sentences that provide unusually balanced prose - but writing with no soul. When used in ad production, such as on Facebook and Google, you get a deadly mix of the two.
I've yet to see a piece of one-prompt AI copy that wasn't recognisable for what it was.
That doesn't necessarily mean it's badly written.
AI Content Isn't Necessarily Bad
Let's be clear: the content that AI creates may not be inherently bad. Unless it's told otherwise, it'll create content that its algorithms think is the best it can do.
This is easily identifiable in the copywriting sphere. Grammar might be perfect, spelling spot on, arguments well-rounded, and it may be factually correct.
However, it will be generic unless effort is made to add feeling, nuance and a point of view grounded in human experience. It will use easily identifiable patterns, and as more people read AI-created content, they start to recognise those patterns. They'll start to see the same rhythm in AI writing, how carefully constructed it is, how neat and balanced.
As an experiment, I just asked Claude and ChatGPT to write an article; it was the same one-shot prompt to each. Once they had done so, I asked them to identify and list the AI copywriting tropes that they had used in the execution. ChatGPT reported it had used 20 obvious constructions; Claude reported 12.
I recognised some that had been highlighted, but I also recognised others that they didn't. Self-diagnosis clearly isn't the answer.
The ones that are most obvious to me were common phrasing we see in AI content, like "actually matters" or short, mic-drop statements like "those are not the same thing" and "the distinction matters". I can't recall seeing that type of construction in marketing copy before AI, and if I see it now, alarm bells ring. Maybe I did, though, just not as often.
Try the experiment yourself: pick a model of your choice and ask it to write an article unrelated to your business; give it as little information as possible. Then, ask the same model, in a new chat, to write an article on a different subject. Compare the two, and see if you can identify the patterns.
We're Learning to Question What We See
When we think of other areas where AI has made huge strides, like video and image creation, we already see evidence that people no longer know what to believe.
Is that reel genuine? Is that news piece truthful? Is that image correct?
Let's be honest, it's getting harder to tell, and some are taking advantage of that.
It's an unfortunate truth that people are gullible; they want to believe what they read or see, especially if it fits their agenda or solves their problem. Therefore, it's easy to produce content that satisfies their thirst for affirmation or plays on their insecurities.
That said, although it's different for each individual, I believe there's a tipping point when people shift from believing to uncertainty in what they see or read, and that develops into downright distrust. From a marketing point of view, that's a problem.
Marketing Depends on Trust
This article is about trust, and more specifically, trust created by content.
When used in the business world, recognisable AI content can be particularly damaging. I see it most often in email, where it's obvious that a reply has been constructed by a machine rather than a human.
Have you ever read an email from someone you know that sounds like they've never met you, or just doesn't sound like them? I have, and I see it more and more, and it makes me feel uncomfortable - am I not worthy of a genuine reply? I get that people are busy, but why risk breaking trust in a medium that is so hard to communicate over anyway?
There's a difference between reading something and thinking it's badly written and reading something and realising that the person supposedly communicating with you probably didn't write it at all. The information might still be perfectly correct, but seeds of doubt have been sown.
If I were to send out a weekly marketing email to clients that sounded distinctly unlike how I normally write, what message does that send? Marketing is supposed to create confidence. If the content being used to achieve that is obviously machine-generated, it risks doing precisely the opposite.
Copywriting Is Still a Craft
Businesses, especially small businesses, rarely see the value in copywriting. When I'm asked to assess a proposal, and I ask who's creating the copy, I'm met with blank stares. "Isn't it the web designer's problem?" No, it's not their problem; it's your problem and one you need to get on top of.
This brings on a mild sense of panic, and rather than trying to think it through, clients naturally search for a solution. Enter AI, stage left. "AI will write the content!" I'm then presented with AI-generated content with no feeling, no understanding of the business, and no connection with their customers.
If you let AI generate your content, your website will be substandard and forgettable. The ability to generate words isn't the same thing as the ability to write good copy.
Good copywriters bring persuasion and psychology, storytelling based on experience (yours or theirs), audience empathy, and they bring a brand voice. AI can provide some of these, but it can't bring empathy, experience or nuance. AI doesn't know your business, your customers, or their problems. It can guess, but that's not the same as knowing.
AI can create wonderfully structured sentences. It can spell correctly, organise an argument and produce vast quantities of content faster than any human copywriter ever could. But those things alone aren't copywriting. Good copywriting is about great communication, connecting with the reader, and leading the reader to a resolution. I don't believe this should be handled solely by a machine.
The problem with AI slop isn't simply that a machine created it; it's that it's recognisable in the content, and once the reader notices that, you're in trouble. They aren't just judging what they've read, they're judging whether it was worth reading. If your business couldn't be bothered to put in the effort, why should the potential customer be bothered to consider your business?