Do you have an AI slop problem? #Newsletter 05

If people think your content is AI, you don’t get judged on intent. You get judged on output.

That’s the part that marketers are missing. AI slop isn’t a problem with generative AI tools. It’s the audience reaction to content that feels mass-produced, generic, and low-commitment.

Once that reaction happens, the audience stops caring about the message and starts questioning the brand. Raptive’s study of 3,000 people found trust drops by 48% when content is suspected to be AI-generated.

You can be using AI responsibly and still lose. You can be writing it all yourself and still lose. It’s impacted by how it reads, not what you used.

This week, I want to dive into the AI slop hype and what it really means.

More People Now Read With Suspicion

Readers aren’t neutral anymore. They’re instantly deciding whether you did the work.

That’s a rational response to what the internet has trained them on over the last 18 months: more templated content, more recycled slop, more posts pushed out because companies ‘can’.

People have learned that a lot of what hits their feed is there because it’s cheap to produce and profitable to distribute, not because it’s worth their time.

So they don’t read in the usual way. They scan for signs it’s been generated by AI, and if it looks that way, they move on. The penalty isn’t debate or disagreement. It’s dismissal.

In Deloitte’s survey, 50% of people say they’re more skeptical of online information than a year ago, and among gen-AI users, 70% say AI content makes it harder to trust what they see.

That’s why generic content is getting hit harder now. It underperforms and also teaches people that brands don’t have anything to say.

What AI Slop Looks Like in Content

Slop isn’t just bad writing. It’s content that avoids committing to anything specific. It says everything but also nothing at the same time.

Readers have learned the pattern language. The common tells are predictable:

• A polished, neutral, bland tone that could belong to anyone
• Advice that applies to every industry
• No details, no opinions, no numbers, no examples
• No brand anecdotes, no decisions, no ‘we tried X and it failed’
• Visuals that look too clean and too generic
• Content that reads like a ChatGPT response

You can produce all of that without AI. AI just makes it cheaper and faster.

The deeper issue is what this does to brand memory. When everything is phrased in the same safe way, the audience can’t tell who it’s for. They may consume it. They won’t retain it. That’s how brands end up forgettable.

How to Test Whether You Have a Slop Problem

The common mistake is asking ‘was AI used to create this?’ The focus should be on asking whether your output is distinguishable and accountable.

Think about these points:

Could anyone publish this tomorrow?

If you took your name or logo off a piece of content and asked people who it belongs to, could they identify it correctly?

If there’s nothing there to distinguish the brand from the rest, your audience will write it off as AI slop

Are you saying anything someone could disagree with?

A lot of slop is true but carries no weight. It lists principles. It avoids taking a position.

If nobody could push back on what you wrote, it probably isn’t specific enough to be useful.

It doesn’t have to be controversial – it just has to spark conversation, engagement, and real thoughts from real people.

Does it include anything that proves you’ve been close to the problem?

Real details: what you tried, what changed, what you cut, what surprised you, what you’d do differently, what you’d never recommend.

If it’s all clean advice with no real experience behind it, it will come across as slop, even when it’s technically correct.

Would you still post it if you were publishing 50% less often?

This is the easiest internal test.

If the honest answer is no, then the post is only there to keep the machine running, not because you had something worth saying. Readers can tell. They don’t need proof.

What are people reacting to: the idea or the formatting?

If the replies are about the hook, the tone, the structure, or ‘this sounds AI,’ your message didn’t work.

That’s not a problem with your audience, but the content output instead.

Using AI Isn’t the Problem

You often hear ‘we don’t publish pure AI content.’ That might be true.

It also doesn’t matter much because readers aren’t looking at your workflow; they’re comparing your writing to everything else they’ve seen, which is mostly the same AI tone.

For example:

• An Ahrefs analysis of nearly 1 million new web pages found 74.2% contained detectable AI-generated content.
• Graphite’s study of 60,000+ new articles reported that by late 2024, more than 50% of new English articles were AI-written.

None of that gives a clean slop percentage. Detection is imperfect, and AI-generated isn’t the same thing as low value.

The key is to understand that people are becoming more familiar with the same generic produced content every day.

What to Change if You Fail These Tests

This isn’t a problem with using generative AI. It’s more of an editorial standards problem.

If you suspect you have an AI slop problem, the changes are:

• Reduce output until you can increase specificity. If you can’t add valuable information, publishing more will mostly train the audience to ignore you.
• Make value density the standard. Every piece should contain something that could not have been produced without your actual context: your numbers, your decisions, your failures, your operating environment.
• Treat generic tone as a defect. If a paragraph could appear in a thousand LinkedIn posts, it shouldn’t make it to the ‘publish’ stage.
• Keep humans responsible for claims. AI can draft, summarize, and structure. Humans should own what’s asserted, what’s excluded, and the intent.

Speed and volume or trust and distinctiveness. Teams can buy speed easily now. Trust is becoming harder to earn.

The Actual Goal Isn’t Less AI. It’s Controlled Output.

The teams doing this well aren’t trying to prove they didn’t use AI. They’re using it to maximize the value they can provide to audiences.

The operating model is simple:

• Use AI for the heavy lifting: outlining, first drafts, variations, restructuring, pulling examples from your own notes, and turning calls into usable text.
• Put one person in charge of the final output whose job is not editing for grammar, but editing for value. Looking at it from an audience > brand perspective.

That second part is where most brands fall down. They optimize AI usage for minimal editing, like it’s a virtue. It isn’t. Minimal editing is how you end up publishing something that is clean but empty.

The time-saving comes from separating tasks properly. AI can produce words fast. It can’t decide what you should be willing to say in public, what you can stand behind, or what your audience will relate to.

If nobody owns those decisions, the output will be slop. That’s the version readers skip.

What Marketers Should Stop Assuming

Stop assuming:

• If it’s accurate, it’s valuable. Accuracy without specificity is still disposable.
• If a human edits it, it won’t read like AI. Light editing often preserves the same generic structure.
• More content protects distribution. Platform policy and audience behavior are both moving against bulk low-value output.
• Nobody notices. They do.

Stop worrying about using AI. Instead, focus on audience perception and value.

That’s what separates unique brand content from AI slop.

If you find value in these emails, please share or subscribe if you haven’t.