AI has fingerprints. Strip them before anyone reads..
Why this mattersPrompt 13 catches what AI gets wrong. This lesson catches what AI gets generic. AI-drafted content has predictable patterns — phrases that signal a chatbot wrote it. Customers can tell. Google can tell. AI detection tools can tell. The slop detection step removes the fingerprints before anyone sees them, and protects your brand voice over time.
Read this once. Sit with it before you answer.
When AI drafts a post for you in 30 seconds, what specifically gives away that AI wrote it, and how do you strip those fingerprints before anyone reads it?
Where you stand right now.
By the end of this lesson, you will be able to:
- 1Identify the five AI fingerprints (em dashes, three-item rhetorical lists, predictable transitions, overused AI vocabulary, generic affirmations).
- 2Apply Prompt 14 (AI Slop Detection Add-on) to every post after fact-checking but before publishing.
- 3Add new entries to your Brand Voice Guardrails page as new patterns surface.
- 4Build the personal style guide that keeps your voice intact over time.
The whole lesson in a few points.
- 01AI has fingerprints: em dashes, three-item rhetorical lists, predictable transitions, overused AI vocabulary, generic affirmations.
- 02AI content detection has matured. Originality.ai, Copyleaks, GPTZero, Sapling spot the patterns. Google's quality systems weight detected AI content less for authority signals.
- 03Prompt 14 flags every fingerprint with suggested replacements. 2 to 3 minutes per post.
- 04Build your Brand Voice Guardrails list over time. 20 to 30 entries within 30 to 60 days.
- 05The goal: AI-drafted content that reads as if you wrote every word yourself.
What AI slop actually looks like.
Fingerprint one. Em dashes. Long horizontal lines connecting clauses. AI loves them. Most humans use commas, periods, or parentheses instead. If you see an em dash in your draft and you do not normally use them, strip it.
Fingerprint two. Three-item rhetorical lists where the last item is the most dramatic. "Faster, smarter, transformative." AI defaults to this structure constantly. Real human writing varies sentence structure naturally.
Fingerprint three. Predictable transitional phrases. "Not just X, but Y." "At the end of the day." "In today's landscape." "The truth is." "Simply put." Once you notice them, you cannot unsee them.
Fingerprint four. Overused AI vocabulary. Delve. Navigate. Foster. Leverage. Streamline. Robust. Cutting-edge. Game-changing. Tailored. Comprehensive. Seamless. Transformative. Empower. Every AI model has its favorite words and they all overuse the same ones.
Fingerprint five. Generic affirmations and filler. "Great question." "Absolutely." "I understand." "I'd be happy to help." Chatbot habits. They do not appear in real human writing unless you are deliberately mimicking customer service language.
Why this matters more in 2026 than in 2024.
AI content detection has matured significantly in the last two years. Tools like Originality.ai, Copyleaks, GPTZero, and Sapling are now used by clients, publishers, schools, and increasingly by Google's own content quality systems. The detection works by spotting predictable AI patterns — the same fingerprints just named.
Content that fails AI detection does not automatically get penalized by Google. But it gets weighted less for authority signals because it reads as low-effort. AI-drafted content that has been stripped of its fingerprints reads as if a human wrote it, which is exactly what Google's quality systems reward.
The slop detection step moves your content from "obviously AI-generated" to "indistinguishably human." Same time investment. Significantly stronger output.
The detection and removal workflow.
Method one. Manual scan. Read the draft. Look for the five fingerprints. Strip them. Replace with your own phrasing. This works once you have done it 10 to 20 times and the patterns become automatic to spot.
Method two. Prompt 14 (AI Slop Detection Add-on). Open your Claude Project. Run the prompt against your draft. The output highlights every fingerprint it finds with suggested replacements. You approve or reject each one. The whole pass takes 2 to 3 minutes per post.
For most students, start with Method two. After two to three weeks, you will start catching the patterns yourself and Method one becomes faster.
The replacement principle is simple. If the prompt flags a phrase you would never say out loud, replace it with how you would actually say the same thing. If the prompt flags a transition you do not normally use, swap it. If it flags a word you do not have in your vocabulary, use the word you would actually use. Read every revised draft aloud. If it sounds like you, publish.
Build your personal style guide.
Open your workbook to the Brand Voice Guardrails page. Add to the list every time you spot a new AI fingerprint that creeps into your drafts.
For example, your list might already include "no fluff," "the truth is," "at the end of the day," "in today's landscape," "let's be real," and several others from your existing brand voice document.
As you keep using AI to draft content, new patterns will emerge. Add them. The list becomes part of your AI prompts so future drafts skip those patterns from the start. Within 30 to 60 days, your list should have 20 to 30 entries. At that point, your AI-drafted content is approaching the quality of content you would write from scratch, in a fraction of the time.
Run Prompt 14 on your next post now.
Run your next post through Prompt 13 (Fact-Check), then through Prompt 14 (AI Slop Detection Add-on) before publishing. Document the slop check on your Publishing Checklist page next to the fact-check note.
Add at least two new entries to your Brand Voice Guardrails page this week — anything you notice slipping into your AI drafts that does not sound like you.
Fact-check first. Slop detection second. Both before publishing.
The last lesson in this module is the automation layer that keeps the entire system running on its own.
The vocabulary that follows you.
- AI fingerprints
- The five predictable patterns that signal AI authorship: em dashes, three-item rhetorical lists, predictable transitions, overused AI vocabulary, generic affirmations.
- Brand Voice Guardrails
- The personal style guide page in your workbook listing the phrases, structures, and words you never want in your content. Grows over time. Becomes input to your future prompts.
- AI content detection tools
- Originality.ai, Copyleaks, GPTZero, Sapling, and others. Spot the same fingerprints by pattern matching. Used by clients, publishers, and increasingly by Google's quality systems.
- Personal style guide
- Your accumulated list of voice rules, phrases to avoid, and structural preferences. 20 to 30 entries within 30 to 60 days for most owners.