How to Not Get Flagged as AI Slop

What is AI slop?

AI slop is AI-generated content published without human thinking, taste, or editing applied. The output goes from the model to the public untouched, and readers can tell.

The tells are consistent: the em dashes, the prose rhythm, the vocabulary. Say "delved" near a marketer in 2026 and watch them flinch. Justin Novak's description of LinkedIn on Session 009 was blunt: everybody is a thought leader now, the content is recognizably machine-written, and the feed has turned into a cesspool nobody wants to read.

“Now everybody is a "thought leader" on LinkedIn, and you can tell it's written by AI, just the em dashes and the prose.”

Justin Novak · 32:00

Can people tell my content was written by AI?

Ian Kilpatrick's answer became the quote of the night: if you are asking, they already can. Ian spots it in YouTube scripts, where he knows who wrote the words even as a human reads them aloud. In his words, this is not a change, it is a tectonic shift.

The failure mode has a public face. The hosts recalled an official who read an AI-written speech aloud and delivered the model's closing sign-off, the "would you like me to" line, to the room. Ian's rule covers it: using AI straight, copying it, and claiming it as your own is where you get in trouble.

“If you're asking that question, people can already tell your marketing is written by AI.”

Ian Kilpatrick · 33:57

What is LinkedIn's AI slop button?

LinkedIn introduced a report-this-as-AI-slop button, letting readers flag machine-written posts the way they would flag spam. Justin called it brilliant: simple, elegant, and a direct economic signal that platforms are starting to penalize slop rather than just tolerate it.

Watermarking is the other half of the squeeze. The question that opened the discussion said Anthropic is adding watermarks to AI content for the EU. Justin was careful to say he does not know the mechanism, but the direction he read from it is clear enough: platforms and regulators both want AI content to be detectable, so betting your brand on nobody noticing is a bad trade.

How do you use AI without producing slop?

Alec Saluga's rule from Session 009: you should be using AI, but it should never be obvious that you are. Everyone on the call uses AI daily, for research, for ideation, for expanding outlines. None of them ship its raw output as their own voice.

The working method is collaboration, not delegation. Draft the outline yourself, let the model expand and refine, then rewrite the delivery so it is unmistakably yours. Alec's point is that delivery is more important than ever: the same AI-assisted speech lands as competence when delivered well and as slop when the sign-off is still attached.

Ian draws one more line worth copying: he builds with AI constantly and uses it for research without disclosure, but he never claims an AI-written blog post as his own writing. The trouble is never the tool. The trouble is the claim.

“It's good to use AI for ideation, for expanding on your ideas, but to completely remove your brain and critical thinking from it is a bad idea.”

Alec Saluga · 35:11

Why does authentic content win?

Slop creates a premium on the real thing. Justin's bet, stated with a timeline: leaning into authenticity wins over the next 1, 3, 5, 10 years, because people are craving human connection.

Authenticity here is a format decision, not a vibe. Every client Justin works with does video, YouTube, in-person meetups, and live online sessions, the formats where a human is visibly present and no watermark question can exist. Session 009 itself is the example: three people, live, answering unscripted questions. At AI4NTP (AI for Non-Techy People), AI runs the distribution behind the scenes while humans make the content. That division of labor is the whole playbook.