How (and why) to avoid creating AI slop

Introduction

At TBWC, we work with brands that want to supercharge their online content and presence, but the onslaught of AI slop is making that increasingly difficult.

Marketing and Editorial teams are under pressure to produce more content, faster than ever. The temptation to use AI is growing daily.

Finally, something that will help you write more in a time-poor and resource-strapped world, right?

No, and that’s the reason human-written content isn’t dead – in my view, as a marketer of 16 years, the complete opposite is true.

The onslaught of ‘AI slop’

2025 has gifted us a new type of internet spam: AI slop. If you’re not familiar with this phrase, it describes “low-quality” content written by AI. 

You may have spotted articles that feel off, even if you can’t say why. Think about it: have you noticed articles that say a lot without going deeper into the problem or topic? Social posts that have more emojis in them than interesting takeaways: the overuse of the contentious em-dash, which results in everyone starting to sound and look the same? (I’ll admit, I’m terrible at this, and I promise I’m a human…)

Don’t get me wrong – by no means is all AI-generated content bad. As I mentioned earlier, if the prompt is sophisticated and there’s a skilled writer in the mix, it can be a worthy addition to the web. It’s allowing people who lack confidence in writing to explore their ideas in new ways, opening up new avenues of creativity, and that’s a great thing.

However, when it’s poorly prompted and just an ‘AI content bot’ chucking in a list of keywords and directly copying and pasting the articles into a CMS, it becomes easy to spot. 

AI Slop examples

Using ChatGPT, let’s take a look at what can happen if the prompt isn’t detailed enough.

Example 1

The prompt:
Write about online strategy

The answer:

“In today’s fast-paced digital world, businesses need to focus on their online presence in order to stay ahead of the competition. One of the most important aspects of digital marketing is having a strong online strategy that works. Businesses of all sizes can benefit from strategies that help them grow online and increase visibility. When it comes to digital success, having the right tools and strategy is key to success.”

Example 2

The prompt:

Write about why content marketing is important.

The answer:

“There are many reasons why content marketing is important. First, it helps businesses grow. Second, it builds trust with audiences. Third, it increases brand awareness. Content marketing is used by many successful companies and has become a key trend in digital marketing today. That’s why investing in content marketing is a smart idea for any business that wants to succeed online.”

The red flags

• Fluff and redundancy: Phrases like “digital world”, “online presence”, and “strategy is key to success” repeat the same idea without adding any value to the reader. 

• No specifics: There are no mentions of what strategies or how they help. It reads like filler.

• Keyword-stuffing tone: I kept the prompts deliberately broad for this example, but it shows that generic prompts produce generic content. There’s no refining or real focus on the topic or intended audience.

• Lazy structure: Reads like it’s checking off a mental list — “First… Second… Third…” — without context or detail.

• No depth or evidence: It states outcomes (e.g. “builds trust”) without backing up how or why.

• Overused buzzwords: “Smart idea for any business,” “key trend,” and “succeed online” are red-flag phrases that signal vague thinking.

• It almost reads like a high school essay, using more words than needed to get your point across. You can almost hear the word count ticking upwards.

• The important question we ask at TBWC – and I ask you now – is “What value is this adding to the conversation online?” If you’re not becoming an interesting voice in the debate you’re trying to enter, it’s not going to benefit you in the long term.

Other tell-tale signs

Of course, the examples above are deliberately poorly prompted to showcase AI slop to you. 

AI content is rapidly becoming more sophisticated, but there are still some tell-tale signs that give away the fact it’s machine-created when you know what you’re looking for.

Repeating the same phrases

AI slop will often overuse transition phrases like “it’s important to”, “in conclusion”, and “despite this”. 

An immediate spot for me is when an article kicks off with “In today’s fast-paced world”… it’s ok AI, we know the world moves quickly, let’s leave it unsaid.

No variety in sentence structure

AI will often pad out its responses in a way that really means nothing when you consider what’s actually being said.

It often uses long, word-salad sentences throughout the entire article, like the example below.

“By leveraging scalable storytelling frameworks across omnichannel ecosystems, brands can holistically optimise synergistic engagement metrics to drive next-gen thought leadership at the intersection of value-driven visibility and actionable authenticity.”

It’s important to remember that AI’s strength is not in long-form content. 

If you find it beyond tempting to use it for this task, be warned: spend a lot of time working on the output, or your user will be instantly disappointed in the surface-level content.

Humanity stripped out

I enjoy reading articles that have a conversational tone, as they have cadence, rich author expertise, and take me on a journey.

AI cannot replicate personal anecdotes or your deep understanding of what drives your target audience. It doesn’t know your customers the way you do. 

It cannot replicate your deep knowledge of their pain points, humour, internal language, or the little cultural nuances that make content just feel right. 

While AI can simulate tone or structure, it can’t draw on lived experience, real conversations with your target audience, or the instincts you’ve honed over years of working on them. 

And that matters, because those details are what build trust, make people feel seen, and ultimately drive action. 

Overuse of bullet points

Now, I’m not saying bullet points don’t have their place, but AI will often focus on delivering information in a quick, easy-to-digest format, as it’s been trained to make things readable.

Again, this is no bad thing, but consider whether you’re giving enough information to action the bullet points, rather than just firing information at the reader.

Think like an editor when reviewing a piece of content, and have these signs to check off in the back of your mind. 

Where should AI be used in the content process?

So, we know not to use AI to entirely write your content, but how should you and your teams use AI as a tool to accelerate your content production? 

After all, most companies can’t afford the luxury of experienced journalists, delving into the heart of their content and creating bespoke, tailored editorial plans chock-full of expertise and flowing, glistening content.

Does that mean everyone else should cease content production? Of course not. AI is a wonderful tool… if used correctly.

How I’ve used AI to help me write this article

Research

Research is my biggest time sink when it comes to creating content. Once I start, I can fall down a 2,000 comment Reddit hole and resurface hours later, none the wiser. 

That’s what I did when planning this piece, heading into useless topics to check I wasn’t missing anything, but feeling like I’d wasted time.

To speed things up, I used Perplexity to get a broad overview of recent updates to Google Search (I am personally a B2B marketer as opposed to an SEO specialist) and to sense-check what others were saying about AI-generated content. 

I found it useful for surfacing expert opinions, like Mark Ritson’s, that I could dig into further (which you’ll see later in the piece). 

I also used ChatGPT’s Deep Research function to explore how marketing and editorial teams are thinking about AI use in content creation. 

But the key is I didn’t take any of it at face value. I double-checked sources, cross-referenced stats, and looked at the original posts or articles that made those claims.

IIf I couldn’t verify it, I didn’t use it. These tools helped me research faster, but the judgment and framing are still mine.

Remember, AI is not necessarily your source of truth, it’s an assistant to do repetitive or large-scale legwork. Always stay sceptical.

Grammar

I’ll hold my hands up: I am terrible at grammar (although, in fairness, I’m far from alone…). Our QA process and spell checkers usually catch these things, but I have a habit of using the passive voice. 

Whenever I spotted myself doing so, I sent it to ChatGPT and asked it to correct the sentence, before going back to my document and making the recommended changes. 

I find this more helpful than a blind copy/paste because I am learning along the way, not as an automatic correction.

It’s another example of using AI as an assistant, and not a source of truth, being able to ask things and get a second opinion when you have doubts or need to explore ideas.

Shortening my sentences

I know this one is a shared pain, but I will stare at a sentence or a paragraph for longer than I need to and still not be able to get it right. 

This is a great use case for AI. I’ve already written the content, I just need to shorten it for the flow of the piece. A caveat here is that I am using it to suggest, not tell. To help fire ideas in my brain, rather than being the source of truth.

So my prompt would be ‘give me three ways to shorten this sentence’ and then I can decide how it should go based on what I want, as otherwise you’re just in danger of injecting ‘micro slop’.

Audit before you automate

Sharing how I’ve used AI here is just an example of how it can be used to assist with the writing process, not replace it entirely. However, I would recommend that teams don’t just follow the generic advice on where to use AI, but instead audit their content workflows first. 

One simple way to start: rate each task in your production workflow using a framework like the one below, which assesses things like expertise, strategic value, and risk. 

AI is brilliant for repetitive, time-consuming jobs where expertise isn’t critical, but the higher the task scores, the more you should keep human hands (and minds) in the mix.

This framework is a selection of steps you should be considering when planning and writing an article, and the questions you should be asking yourself at every step in the article production process:

Scoring

0–15 – Strong AI fit (ideal for automation)

16–25 – Hybrid use (AI-assisted with human oversight)

26+ – Human-led (keep AI use to a minimum)

The above exercise can be beneficial to help you determine, as a team or company, when AI usage is appropriate. 

Now, a few hours is not a lot of time, but I am one person, writing one article this week.  For a team of five producing dozens of articles, those saved hours quickly become a big advantage.

So…should I stop using AI if I want to make an impact?

No, not at all. Used correctly, AI is a superpower for writing and content creation.

But knowing what to look for when it comes to AI content is much more than just catching a blog that might come from your team or freelancers. 

It’s about protecting your brand and why your audience comes to you for content in the first place. 

The brands that rely on AI to write their articles to meet lofty output goals (often set by managers who aren’t paying attention to workloads) are all about to look and sound the same. 

That doesn’t mean that marketing and content teams should avoid using AI (quite the opposite), but AI should be used as a tool, not as a crutch. 

If you unknowingly or deliberately publish low-value content (AI-produced or otherwise), you risk damaging your search authority and eroding audience trust. 

Google and AI slop

Google is already catching up and penalising sites that rely on AI-produced content. 

In the “Helpful Content Update”, Google prioritises content “written primarily for SEO, not humans,” and in the March 2024 core update, which specifically targeted and down-ranked “copycat” or mass-produced content, much of it created using AI tools. 

Over 1,400 sites reportedly received manual penalties following this update, with many sites experiencing dramatic drops in rankings or complete removal from search results for relying on generic, low-value AI content (thanks, Perplexity!).

With the emphasis on E-E-A-T, this is just one area where AI cannot compete. Google knows when copy is written by AI, and your team should too. 

There is so much fear around AI being the death of content writers, but to me, there is an opportunity. 

Brands looking to elevate themselves above the AI slop are about to massively benefit from.

Brands that invest in their human writers, who understand their audience and follow E-E-A-T principles, are about to stand out in a sea of AI-generated fluff.

And I’m not the only one saying it.

Legendary brand consultant and marketing professor Mark Ritson has been vocal about the growing flood of synthetic, low-value content. 

It’s the making of a “zombie internet” — one where recycled, generic material drowns out originality.

His stance is clear: AI is useful, even powerful, for automating repetitive, high-volume tasks like social posts or comments. 

But when it comes to strategy, brand-building, or creativity, it’s just a tool, not a replacement.

In other words, the opportunity isn’t to be replaced by AI. It’s to be freed up by it, so your human creativity has more room to breathe.