Media
2 items
Description
Evaluates LinkedIn feed posts for likelihood of being AI-generated using heuristic analysis. A Chrome extension that analyzes LinkedIn feed posts and scores them for likelihood of being AI-generated using heuristic pattern matching. What It Does As you scroll your LinkedIn feed, the extension automatically analyzes each post and adds a small colored badge showing an AI probability score (0–100%). Click any badge to see a detailed breakdown of which signals were detected. | Score | Color | Meaning | |-------|--------|------------------------| | 70–100 | 🔴 Red | Very likely AI-generated | | 50–69 | 🟠 Orange | Likely AI-generated | | 30–49 | 🟡 Yellow | Possibly AI-assisted | | 15–29 | 🟢 Light green | Mild AI signals | | 0–14 | 🟢 Green | Likely human-written | Detection Signals The extension checks for: AI Phrases — Overused LLM phrases like "in today's rapidly evolving", "let's unpack", "delve into", "paradigm shift", etc. Structural Patterns — Numbered lists, emoji-heavy formatting, hook→list→CTA formulas, one-sentence-per-line "bro-poem" style Sentence Uniformity — AI text tends to have suspiciously uniform sentence lengths Hedging & Filler — Excessive qualifiers like "it is worth noting", "broadly speaking" Engagement Bait — "Agree?", "Thoughts?", "Tag someone who...", "Repost if..." Repetitive Starters — AI often begins multiple sentences with the same phrase Limitations This is a heuristic tool, not a machine learning classifier. It looks for stylistic patterns commonly associated with LLM-generated text. That means: - Some human writers who use corporate buzzwords will score high - Well-edited AI text that avoids common patterns will score low - It's best treated as a "sniff test," not a verdict