osint
The 10-Minute Truth: Using AI to Spot Fake LinkedIn Profiles
LinkedIn has a billion profiles, but how many are actually real? I ran an OSINT audit on my own profile to see what AI-powered due diligence reveals about professional credibility.
The Billion-Profile Blind Spot
Everyone sees the high-level titles: "VP of AI Strategy," dozens of certifications, board seats. But in a world where anyone can claim anything, the real question is: how much of it is true? Most of us take these profiles at face value because deep verification is perceived as too time-consuming.
I decided to flip the lens and run a structured OSINT (Open Source Intelligence) analysis on my own profile to see what a professional investigator would actually find.
The Experiment: A Due-Diligence Audit
I used a structured prompt to perform a full risk review. This wasn't just a surface-level glance; it involved cross-referencing App Store listings, employer records, domain ownership, and cross-platform handles. Every claim was scored against verifiable evidence.
Three Surprising Findings from the Data
The results highlighted some fascinating patterns in how "authenticity" is measured today:
- Scam Pattern Triggers: A polished headline like "Open to partners" often triggers the same detectors as scam or MLM profiles. Context matters to humans, but algorithms flag the pattern first.
- The AI Language Flag: High-quality English from a non-native speaker can be flagged as potentially AI-generated. In a global tech market, 20 years of international experience can look exactly like a GPT output to a risk-scoring bot.
- The Authenticity of Inconsistency: We often crave uniform growth, but humans are messy. A post with high impressions compared to a low follower count is actually more authentic than perfectly uniform stats. Bots are consistent; humans aren't.
The New Standard for Due Diligence
My profile scored an 18/100 risk—low, but not zero, largely due to incomplete data. The takeaway is simple: this analysis took only 10 minutes and a well-structured prompt.
Before your next hire, partnership, or investment, run the check. The tools exist to separate the signal from the noise.
Want the method? DM me for the scraping method and the research prompt, or find it in the database at PUCO.app.