AI Resource Lab

I Used AI to Find My Perfect Doppelgänger (Then Hired Them to Replace Me)

ai doppelganger
Jeremy Gallimore AI

Jeremy Gallimore

Experience Designer | Visual Storyteller | AI Innovator

Some of the links in this blog post are affiliate links, which means I may earn a commission if you make a purchase through these links at no additional cost to you.

The Day I Met Myself

Last Tuesday, I had a conversation with myself. Not in the mirror—not in a daydream—but on a Zoom call where my digital twin nodded along to my thoughts, laughed at my jokes, and even scratched its nose the exact way I do. By day three, this AI clone was interviewing for my job.

This wasn’t science fiction. This was a Tuesday.

What started as a curiosity—“What if I could outsource my existence?”—turned into a full-blown identity heist. Over 72 hours, I replaced myself with two doppelgängers: one digital, one flesh-and-blood. The experiment revealed just how fragile our concept of identity has become in the AI era.

The Process: Building a Digital + Physical Doppleganger

I began with the easiest part: creating an AI clone. Using HeyGen, I fed 30 minutes of my podcast rants into the system. Within hours, it had mastered my speech patterns—the way I overuse the word “literally,” the awkward pauses before punchlines, even the sarcastic lilt I deploy when explaining crypto to normies.

Then I plugged it into Synthesia, which animated my clone with unsettling precision. The first time I watched it back, my stomach dropped. There I was—or at least, something that looked, sounded, and moved like me—ranting about Web3 jargon without any input from my actual brain.

The real test came when I let it loose in the wild. I had my digital twin:

  • Record a podcast intro (my producer didn’t notice)

  • Answer client emails (they replied “Thanks for the quick turnaround!”)

  • Even flirt with my girlfriend over text (“Wait… this doesn’t sound like you,” she said—but only after three exchanges)

The most disturbing moment? Waking up at 3 AM to find my clone had joined a Twitter Spaces debate while I slept—and was winning.

    Finding My Human Lookalike

    But a digital clone felt like cheating. I wanted to see if a real person could pass as me.

    Enter PimEyes, the facial recognition tool that scours the internet for your doppelgänger. I uploaded a series of unflattering selfies—morning bedhead, mid-sneeze, that one angle where my nose looks huge—and waited.

    The algorithm spat out Eduardo, a 34-year-old line cook from Lisbon. The resemblance was uncanny: same crooked smile, same receding hairline, same habit of raising one eyebrow when skeptical. I slid into his DMs with: “Ever been told you look like a broke tech bro?”

    His response: “Every damn day.”

    We struck a deal: For $200 and a case of Portuguese wine, Eduardo would spend a weekend being me.

    The Swap

    I gave Eduardo three tasks:

    • Wear my signature black hoodie to a local tech meetup
    • Sign copies of my book at a café (“Your signature sucks,” he informed me)
    • FaceTime my mother (“You sound sick,” she said—but didn’t question it)

    Meanwhile, my AI clone handled my professional life:

    • Recorded a podcast episode (my co-host praised “how present” I seemed)
    • Negotiated a contract (the client signed without hesitation)
    • Even attended a Zoom board meeting (where it nodded sagely at all the right moments)

    The results? Nobody noticed. Not my colleagues. Not my friends. Not even my mother.

      Why This Should Terrify You

      This experiment revealed three uncomfortable truths.

       

      • Identity is now hackable
      • For less than $500, anyone can create a “good enough” version of you. Scammers don’t need perfect replicas—just plausible ones.

      • We’re primed to trust fakes
        People wanted to believe Eduardo and my AI clone were me. Confirmation bias is a hell of a drug.

      • Authenticity is a performance
        Watching Eduardo mimic my mannerisms made me wonder: How much of “me” is just a series of repeatable patterns?

      The Tools That Made It Possible

      • HeyGen/Synthesia: For creating AI clones that can pass casual scrutiny

      • PimEyes: For finding your real-world doppelgänger (use responsibly)

      • DeepSeek: We used our own AI to analyze the ethical implications—turns out “Is this illegal?” is a complicated question


      Try It Yourself (At Your Own Risk)

       

      Want to see your own doppelgänger? Here’s how:

      • Use Elevenlabs to clone your voice

      • Go To Synthesia to create your doppelganger
      • Run your face through PimEyes

      • Screenshot the results and tag me—I’ll feature the creepiest finds

      Because the real question isn’t “Can you spot a fake?”—it’s “How many have already slipped by you?”

      About the Author

      Jeremy Gallimore is a leading voice in AI reliability, blending technical expertise, investigative analysis, and UX design to expose AI vulnerabilities and shape industry standards. As an author, researcher, and technology strategist, he transforms complex data into actionable insights, ensuring businesses and innovators deploy AI with transparency, trust, and confidence.

      Who We Are

      AI Resource Lab is the industry standard for AI reliability benchmarking, exposing critical flaws in today’s leading AI models before they reach production. Through adversarial stress-testing, forensic failure analysis, and real-world performance audits, we uncover the hallucination rates, security vulnerabilities, and systemic biases hidden beneath marketing hype. With 15,000+ documented AI failures and proprietary jailbreak techniques that bypass 82% of security guardrails, we deliver unmatched transparency—helping businesses, researchers, and enterprises make smarter, risk-free AI decisions. Forget vague promises—our data speaks for itself.

      Follow us for insights and updates: YouTube | LinkedIn | Medium:

      Related Articles

      ECOVACS Deebot X2 OMNI Review

      ECOVACS Deebot X2 OMNI Review

        The ECOVACS DEEBOT X2 OMNI is a next-generation robotic vacuum and mop, designed for ultimate precision cleaning. Featuring a unique square design, powerful 8000Pa suction, advanced AI-powered obstacle navigation, and an all-in-one OMNI station, this device...