"The cover letters all sound the same," she said. "Perfect grammar. Perfect structure. Three paragraphs. Opens with enthusiasm for the company mission, transitions to relevant experience, closes with a forward-looking statement about contributing to the team. It's like they all went to the same writing class."
They didn't go to a writing class. They went to ChatGPT.
The numbers are stark
A 2025 survey found that 53% of hiring managers have rejected candidates they suspected of using AI to write their entire application. Not some. Not a few. More than half.
At the same time, the volume of applications has exploded. It's never been easier to blast out 50 applications in an afternoon when AI does the writing for you. Companies are seeing 2–3x more applications per role than two years ago, and a significant chunk of them are clearly machine-generated.
This has created a strange inversion. In 2023, a polished, articulate cover letter was a positive signal — it suggested the candidate was thoughtful and professional. In 2026, that same level of polish triggers suspicion. Perfection has become generic.
What AI-generated applications look like
Hiring managers and recruiters have gotten surprisingly good at spotting them. The tells aren't about grammar or formatting — they're about voice and specificity. Or rather, the absence of both.
Common patterns they flag:
- Smooth but empty. The writing flows well but says nothing concrete. Lots of "I'm passionate about leveraging innovative solutions" without any specific example of doing so.
- Uniform enthusiasm. The candidate is equally excited about the company's mission, the team culture, the growth trajectory, and the tech stack. Real humans have opinions and priorities. They care about some things more than others.
- No rough edges. Every sentence is grammatically perfect. Every paragraph transitions smoothly. Real writing has personality — slightly awkward phrasing, a joke that's more clever than funny, a sentence that runs a bit long because the person had a lot to say about something they care about.
- Suspiciously comprehensive. The cover letter addresses every single requirement in the job posting, in order, almost as if someone copy-pasted the JD into a prompt.
The deepest irony: AI-generated applications often look more professional than human-written ones. But professionalism without personality is exactly what recruiters are now trained to distrust.
The right way to use AI in your job search
I'm not arguing that you should avoid AI entirely. That would be ridiculous — it's a genuinely useful tool. But there's a difference between using AI as an assistant and using it as a ghostwriter.
Use AI for:
- Research. "Summarize this company's recent product launches and funding history" is a great prompt. It saves you 30 minutes of Googling and gives you material to reference specifically in your application.
- Brainstorming. "Here's my background. What angles might be most relevant for a role focused on developer tools?" This helps you figure out which experiences to highlight.
- Editing. Paste in your draft and ask for feedback on clarity, typos, or structure. Let it tighten your prose — but start from your prose.
- Keyword matching. "Compare my resume to this job description and identify gaps." This is genuinely useful for getting past ATS systems.
Don't use AI for:
- Writing the whole thing from scratch. "Write me a cover letter for this role" produces the exact output that hiring managers are now filtering for.
- Generating your "story." Why you want this job, what you've built, what you care about — this needs to come from you. Not because AI can't write it, but because AI writes it the same way for everyone.
What actually stands out now
In the sea of polished AI prose, the applications that get attention are the ones that feel human. That sounds vague, so here's what it looks like in practice:
Specificity over polish. "I spent last weekend trying to use your API to build a Slack integration, and the OAuth flow had me stuck for two hours. Happy to talk about what tripped me up and how I eventually got it working." A recruiter reads that and thinks: this person actually used our product. You can't fake that with a prompt.
Opinions over enthusiasm. "I think your pricing page is underselling the enterprise tier — the feature comparison doesn't surface the audit logging and SSO, which is what enterprises actually buy on." That's a candidate with judgment, not just excitement.
Honesty over comprehensiveness. "I haven't worked with Rust in production, but I've been writing it on side projects for a year and I'm confident I can ramp quickly. Here's a repo." Admitting a gap and then showing how you're closing it is more credible than claiming you check every box.
The meta-irony
We're in a weird moment. AI makes it trivially easy to produce polished job applications, which makes polished job applications less valuable, which means the candidates who stand out are the ones who sound most like... themselves.
The best job applications in 2026 aren't the most professional ones. They're the ones where you can hear an actual person thinking. If you use AI, use it to save time on the boring parts — the research, the formatting, the keyword optimization. Then write the parts that matter in your own voice, rough edges and all.
A slightly imperfect cover letter that clearly came from a human who cares will outperform a flawless one that could have come from anyone.
