Is it Possible to Use AI Without Making AI Slop?

Key Points

• Kathleen Booth's argument that marketers have hit rock bottom on AI slop is not wrong, and we feel a bit seen on the framing of using it as an addiction.
• But when does slop happen? Usually for teams using AI as an author, not as an assistant.
• AI authorship with no original source material, no oversight, and no editorial gate is what creates slop.

Yes the AI Slop Reckoning Is Real

Kathleen Booth is right that the industry has hit a bottom of sorts. We'd add one thought on top: the point of the detectors coming for you aren't to measure whether AI touched your content. They're really trying to measure whether anything HUMAN did. That's the underlying goal at the end of the day.

That's a different problem with a different fix, and most teams are about to solve for the wrong one.

Watch what happens over the next quarter. Plenty of marketing teams will respond to watermarking by shopping for evasion. Booth already saw it happening in her feed, marketers trading Claude Skills that supposedly strip AI language patterns. That's a team spending time and money to defeat a detector instead of spending it on the making the content more human from the start instead.

Two Announcements, One Week

Booth's newsletter, Code Meets Creed, connects two announcements that landed within days of each other. Claude is rolling out AI watermarking on all generated content globally. Clay published an AI writing policy that got picked up across tech press. Both follow on the heels of announcements by Substack and LinkedIn that they are working on AI slop detectors of their own.

Booth revived the forgotten art-form of writing the piece entirely herself, with one quoted exception. She also admitted she's contributed to the problem, recently, out of some mix of laziness and speed and the belief that more content equals better content. (Can't say we don't understand that one a bit ourselves!)

Essential Background Reading:

• Content Conductor: AI Content Production for Brands: An overview of how the platform builds articles from a brand's actual source material instead of a keyword and a prayer.

AI as an Assistant vs. an Author

Booth's central distinction is that AI works as an assistant and editor, and fails as an author. We agree with the second half and wholeheartedly believe there's a way to at least make AI a co-author without creating slop.

AI-as-author isn't inherently the problem. AI as an author-with-nothing-to-say is. A model given a keyword and no source material will produce the statistical average of everything ever written about that keyword, exactly as Booth describes with her 100,000-people-finishing-a-sentence analogy. The limitation there belongs to the brief.

Give the same model a sales call transcript, a customer postmortem, a research doc your team just finished, and a set of highly detailed rules about what your brand should and shouldn't say, and the output stops being so average, because the inputs stopped being public. That in no way negates that every article should have heavy human injection in review and editing as well.

We also don't have a clean answer for the disclosure question Booth raises, and are just as fascinated by where this one will go. If the source material is yours and the review is rigorous, does a watermark on the draft mean anything? We think the answer is really dependent on if the content is useful and solves problems for the reader (as the litmus test has always been). While nobody has a good version of this answer yet, we refuse to believe that helpful content won't still be rewarded.

Related Content:

• Content Conductor vs. Jasper: Where a prompt-based tool and a source-material-based pipeline actually diverge, feature by feature.
• Content Conductor vs. Copy.ai: A direct look at output quality and brand voice fidelity between the two approaches.

What Detection Measures

Because this is an article about the launch of new AI detection methods, we do need to discuss them.

Next Steps:

• Agency vs. Content Conductor: What the same accountable-review structure costs when an agency runs it versus when your team does.

What to Do Before the Next Policy Drops

The teams that come out of this fine will be the ones who are still creating great content, and that means it starts from a human's brain. Evasion prompts won't be the solution. We believe better, more useful, more interesting content will be the solution (regardless of if AI helped a bit, and regardless of if those benefiting from it know that AI was used).

Start here:

• Write the AI policy now, publicly: Clay's got attention because it existed. Yours doesn't need to be clever, it needs to be specific about where AI is used and where it isn't.
• Audit your inputs, not your outputs: Pull ten recent articles and ask what real source material each one started from. If the answer is "a keyword and a competitor's H2s," you found your slop.
• Name one accountable reviewer per piece: Not a committee, one person who read it and would defend it.
• Kill the volume target that's driving this: If your content calendar demands twelve posts a month and you have material for four, the other eight are the mess everyone's complaining about.

Stop rewriting your prompts to sound less like AI. It's the fourteenth version of the same prompt and it isn't working, because the problem was never the prompt, it was the lack of unique input. Fix where your content starts from.

Where We Land

Booth ends on intervention and accountability, and we agree with every bit of that. So how do you turn an intention into a sustainable practice? What makes an intention survive is repeatable process and structure: a defined input, a named reviewer, and a gate before anything goes out under your name.

Content Conductor™ was built on that assumption. Real source material in, multi-stage pipeline through Draft, SEO pass, Editorial pass, AI-optimization pass, Schema markup, and CMS staging, then staged for your review rather than published. You are the conductor. AI is the orchestra.