When 80 Percent of Applications Are AI-Generated: What It Means for Your Selection Process

Marcus Fischer
Jul 29
3 min read

In his column for Personalwirtschaft, my highly esteemed colleague Marcus Merheim writes about a recruiter who received 340 applications for a single job posting — 280 of them obviously written with AI. No typos, no odd phrasing, no rough edges – just consistently competent-sounding text that could just as easily have gone out to a hundred other companies. Roughly 80 percent AI share in a single applicant pool. Not an isolated case, but a preview of what is becoming the new normal.

From Exception to the Rule

The numbers confirm the impression. According to the Softgarden study „AI Meets Recruiting 2026,“ the share of applicants actively using AI for their application has more than tripled since 2023 – now standing at 43.2 percent for cover letters alone. Counting everyone who is at least considering using AI, the figure rises to 73.5 percent of applicants. On the other side of the table, the picture looks different: 68 percent of companies still don't use AI in recruiting at all. An imbalance emerges – applicants are upgrading their tools while many recruiting teams are still screening like it's 2019.

The result: more applications, more uniformity, less differentiation. And a selection process that was built for a different application volume and a different level of text quality.

Why Screening Alone Won't Solve the Problem

The obvious reflex – detecting and filtering out AI-generated applications – falls short, for two reasons:

First, AI-generated text is hard to reliably detect, especially in carefully revised cover letters. Automated filtering here produces false positives and risks losing good candidates for purely formal reasons.

Second, a blanket rejection of AI-assisted applications sends exactly the wrong message to the very people you want to win over: competent, tech-savvy candidates who used a tool that has by now become as unremarkable as spell-check.

Screening against AI is an arms race in which, in the end, AI faces off against AI – and the two people who actually want to find each other get lost in between.

The truly relevant question is a different one anyway: not whether a cover letter was written by a human, but whether this person can do the job.

Four Concrete Adjustments for Your Selection Process

Anyone who wants to break out of the negative spiral doesn't need better AI detection, but a selection process that's less vulnerable to AI noise:

Devalue or drop the cover letter.

If the cover letter was never a reliable performance indicator to begin with, now is the moment to remove it from evaluation – or replace it with a short, task-related question that can't be generated in seconds.

Screen for skills, not documents.

Structured short assessments, concrete work samples, or case tasks reveal what a résumé can't. The effort per candidate rises slightly, but so does the hit rate – and pre-selection becomes more robust against sheer volumes of text.

Move structured interviews earlier in the process.

A short, standardized first conversation separates signal from noise more reliably than any prompt-generated application document.

Use active sourcing deliberately against the flood.

Where inbound channels are swamped with AI-generated applications, the lever shifts toward proactively approaching candidates who actually fit – more effort, but far less susceptible to dilution by sheer mass.

None of this happens overnight. But every step reduces dependence on a document that never measured what actually matters.

What the AI Application Flood Does to Your Employer Brand

The truly underestimated part of the story: how a company handles the AI application wave has itself become a signal – to the outside world. The employer brand isn't built only on the careers page, but above all in the process: in response time, tone, and the appreciation applicants actually experience. Weeks of silence or an automated blanket rejection over suspected AI use gets told and retold just as much as a positive experience – only faster and louder, on Kununu, LinkedIn, or among friends.

Anyone who, faced with rising application numbers, only tightens the screening screws without rethinking the application experience itself risks more than a few bad hires. They risk the reputation that shapes every future hiring round. The AI application flood is therefore less a technology problem than a stress test for a company's own employer-branding substance – and for whether processes still look like they stand for something, even under volume pressure, rather than pure defensiveness.

80 percent AI-generated applications aren't just a warning sign for the future – they're a snapshot of the present. Anyone who merely upgrades their ability to detect and filter out AI applications is only prolonging the arms race. Anyone who instead redesigns the selection process so that real skills become visible – while keeping an eye on the application experience – turns the flood into a competitive advantage.