The Resume Is Dead. Long Live the Evidence.
Generative AI has made the resume easier to fake than ever. But the real problem isn't cheating — it's that the wrong artifact was at the center of hiring for decades. Here's what replaces it.
Seventy-one percent of candidates are now using AI to write their resumes. Not polish — write. And honestly, it's hard to blame them. The resume was always a performance: a curated summary of someone's best-lit moments, optimized for a screener who'd spend eight seconds on it.
AI just made that performance cheaper and faster than ever before.
The response from most hiring teams has been to panic, add more steps to the process, and treat any sign of AI assistance as evidence of dishonesty. That's the wrong frame. The question isn't whether candidates are using AI. They are. So are your recruiters. The question is whether you've built a hiring process that can actually tell the difference between someone who can do the job and someone who can describe doing the job.
Most haven't.
The artifact was always the problem
Here's the uncomfortable truth: the resume was never a good signal. It was a convenient one. Standardized, easy to stack-rank, easy to reject without calling anyone. The degree at the top served as a proxy for grit and intelligence — a shortcut that saved time and introduced decades of socioeconomic filtering as a side effect.
Skills-based hiring has been eating through that assumption for years. Google dropped degree requirements for most roles in 2016. By 2026, 81% of tech employers have followed. Not out of altruism, but because the data is clear: a take-home project or a live problem-solving session predicts job performance better than a credential from a school someone attended fifteen years ago.
AI didn't break hiring. It revealed that hiring was already broken. It just made the cracks too obvious to ignore.
What "both sides are AI-enabled" actually means
HireVue's 2026 hiring report puts it plainly: hiring is no longer a one-sided evaluation. Candidates use AI to write resumes, draft cover letters, prep for interviews. Hiring teams use AI to screen, score, and summarize. The result is a kind of arms race where the signal-to-noise ratio on both sides has collapsed.
Nearly every hiring team — 99% in that same report — considers AI-assisted applications a problem. But 62% still call it "smart" when candidates use AI tools. The cognitive dissonance is real, and it points at something important: the problem isn't the tool. It's the test.
If your hiring process can be gamed by a well-prompted language model, your hiring process was testing the wrong things. A candidate who uses AI to produce a perfect answer to "Tell me about a time you demonstrated leadership" isn't cheating a good process. They're exposing a bad one.
What actually works
The answer isn't to ban AI from the application process or add lie-detector rounds to interviews. It's to change what you're measuring.
Behavioral interviewing with specific, non-generic prompts. Not "describe a challenge you overcame" — that's scriptable. Instead: "Walk me through the last time you changed your mind about something important in your work. What was the evidence?" Specificity is AI-resistant. Personal context is AI-resistant. Contradiction is AI-resistant.
Hands-on tasks that mirror the actual job. A 90-minute paid trial or a realistic work sample tells you more than six rounds of competency interviews. Not a case study designed by McKinsey — an actual slice of the work, evaluated against actual criteria.
Structured interviews with consistent rubrics. Seventy percent of forward-thinking hiring teams have moved here, and the data is straightforward: when every candidate faces the same questions evaluated against the same criteria, bias has fewer places to hide and performance prediction improves.
The goal is to create space for candidates to show how they work — not just what they've done.
What this means for founders
If you're a founder hiring your first ten people, you don't have the luxury of a broken process. Bad early hires compound. The resume is a starting point, not a verdict. Use it to get to the phone screen; use the phone screen to get to the evidence.
Ask for a work sample. Run a structured conversation. Pay candidates for their time on longer tasks — the ones who drop out when you start asking for real work were probably never going to do real work for you anyway.
And stop treating AI-assisted applications as the enemy. The candidates who know how to use AI effectively are, in many cases, exactly who you want. The tell isn't whether they used the tool. It's whether they can defend the output.
That's the filter. Build your process around it.
Jonathan Bouaziz is the founder of Umamy, an AI-powered hiring platform for founders who want to hire well the first time.