
The Hiring Reset: The rise of fake work
Welcome to The Hiring Reset, where we rethink how hiring actually works in the age of AI.
Håkon
2025-08-05
This week, we're diving into a new kind of problem: fake work. Not just fake resumes or inflated job titles – entire fake companies, fake teams, fake conversations. AI didn't create this mess, but it's making it exponentially easier to pull off.
As everything becomes easier to fake, everything real becomes more valuable.
Here's what's happening.
When AI Makes Everything Look Real – Except Trust
Picture this: A freelance technical manager logs into Slack for his first day at a new company. It's his birthday, and the team seems welcoming. Messages flow. Code snippets get shared. Everything feels... off.
The responses cometoo quickly. The English istoo polished. Every conversation reads like ChatGPT cosplaying as a busy developer.
Because it was.
The entire "team" was AI-generated. The endgame? Getting him to download malware disguised as project files.
This isn't just another phishing story. It's something bigger:In the age of AI, trust has become the scarcest resource in hiring.
We're witnessing an authenticity apocalypse:
🤖 Fake Candidates
- Resumes with AI-perfected bullet points and suspiciously consistent career progressions
- "Developers" who ace text interviews but mysteriously lose internet when screen-sharing starts
- Portfolio websites spawned overnight with GPT-generated case studies
🏢 Fake Jobs and Companies
- Startups manufacturing traction with bot-generated metrics
- "Teams" that exist only in carefully orchestrated Slack performances
- Job postings for positions that were never real, just data harvesting operations
💬 Fake Conversations
- Recruiters copy-pasting AI feedback to appear thoughtful
- Reference calls where both parties might be reading scripts
- Interview answers so polished they've lost all human texture
But here's what everyone's missing: As everything becomes easier to fake, everything real becomes more valuable.
Think about it. When anyone can generate a perfect resume in 30 seconds, what stands out?
- The candidate who stumbles explaining their messiest project
- The reference who laughs, remembering that time you broke production
- The portfolio with actual commit histories and embarrassing early work
We're entering the era of the Authenticity Premium.
At Vouch, we've started to bake this in, for instance, by letting teams request peer references upfront, bringing referrals higher on the agenda, and nudge for more human touches in the interview stages. Not because AI is bad (on the contrary, since we are building an actual AI-native hiring system), but because you need better contrast. You need real.
In an age of polished fakes, unpolished realness is your moat.
🔍 Signal of the Week: The Slack Simulation Scam
What happened: Kjartan Manvelyan's viral LinkedIn post exposed a fake company using GPT to simulate an entire development team on Slack. Daily standups, code reviews, water cooler chat – all artificially generated to trick a new hire into downloading malware.
Why it matters:This isn't just a security threat. It's a preview of how sophisticated fake work environments can become. If a scammer can simulate a convincing team, what else can they fake? In remote work, presence ≠ authenticity. The new rule: verify through calls, mutual connections, and trust your gut when something feels too perfect.
✍️ What We're Seeing
It's getting harder to be an applicant, too.
We just spoke to a candidate who’d applied to 130 roles. Every CV was tailored (with ChatGPT’s help). He heard back from a fraction, landed three interviews.
Now he's hiring, and he gets it. His inbox filled up in hours. He's now switching to Vouch to help cut through the noise earlier; for him, his hiring managers, and candidates.
He's still doing 15-minute calls with the most promising applicants to confirm the basics: are they legit, are they real? But our AI now filters for clarity and context before that stage, so he can spend more time where it matters.
His takeaway is that authenticity is now the currency; on both sides. Your job post needs to feel real (even with help of AI). Candidates need real signals (not just polish). The built-in reference request in Vouch is already surfacing that – a reminder that trust starts before the interview.
🔮 The Hiring Shift
How smart teams are adapting to the age of fake work:
1. Prioritise Proof of Work
- Live coding sessions over take-home tests
- Screen recordings of actual projects
- GitHub histories that show learning curves, not just highlights
- Ask candidates to walk through their actual work – messy parts included
2. Build Your Trust Network
- Peer references requested upfront (built into Vouch by default)
- Warm introductions weighted 10x higher than cold applications
- "Who would vouch for you?" as the new first question
- Actually call that third reference – the one they hesitated to list
3. Design for Authenticity
- Write job posts like a human would read them
- Include specific, weird details about the role and team
- Use unscripted video or podcasts in your job ads (supported by Vouch)
- Create "day in the life" content that's actually honest
4. Value Rough Edges
- The typo that proves a human wrote this
- The GitHub profile with abandoned side projects
- The honest "I don't know, but here's how I'd figure it out"
- Candidates who ask unexpected questions
5. Add Human Checkpoints
- Quick "proof of human" calls before deep interviews
- Technical discussions that go off-script
- Reference conversations that dig into specifics
- Real-time problem solving, not rehearsed answers
The bottom line: Stop optimising for polish. Start optimising for genuine.
In the age of perfect fakes, your rough edges aren't bugs. They're features.
About Vouch – your hiring assistant
Vouch is your AI-native hiring assistant. Imagine a recruitment system that actually understands your company, your roles, and your candidates – not just tracks them. We help modern teams scope roles, find better candidates, and make faster, smarter hiring decisions. Everything you'd expect from an ATS is included, but the real value is in how it’s used. Learn more here.