Job seekers are rewriting their applications so they don't sound like AI

Disclosure: I'm the founder of Resumize, a resume tool that uses AI.

Last week Greenhouse published its 2026 Future of Hiring Candidate Report, a survey of more than 2,800 job seekers in the U.S., U.K., Ireland, Germany and Australia. 77% of U.S. job seekers used AI in their search in the past year. 60% changed, or thought about changing, how they write so their applications wouldn't be mistaken for AI-generated content, and 39% actually rewrote one.

So most people use AI, and most of them now worry it shows. Sharawn Tipton, Greenhouse's Chief People Officer, said: "For the first time I can remember, neither side thinks hiring is working."

I spend a lot of time reading resumes, many of them written with AI, and I think the worry is fair. Where people go wrong is hunting for single "AI words" to delete. What gives a resume away is that every bullet sounds like every other candidate's.

What makes a resume read as AI-written

• Verbs nobody uses out loud, like "spearheaded", "leveraged", "orchestrated" and "drove synergies". People say "led", "used", "ran" and "built".
• Numbers with nothing behind them. "Improved efficiency by 30%" doesn't say what was measured or how, so a recruiter can't check it and an interviewer will ask.
• The same sentence shape in every bullet: verb, task, "resulting in" outcome, eight times in a row. Real work isn't that tidy.
• A summary that could describe anyone. "Results-driven professional passionate about delivering impact" would still read the same with another candidate's name on top.

A before and after

This example is made up for this post.

Before: Spearheaded the optimization of customer onboarding processes, leveraging data-driven insights to drive a 35% increase in efficiency.

After: Rewrote our onboarding checklist after noticing most support tickets came from the same two setup steps. New customers now finish setup in one call instead of three.

The second one names the problem, says what you did, and gives a result you could explain in an interview without notes.

What not to do

The same report found that some candidates trying to get past automated screening add skills or keywords they don't have, and some hide text meant for AI reviewers. I'd avoid both. A skill you don't have tends to come up in the first technical interview.

Hidden text can show up the moment someone pastes your resume into a plain-text field, and after that the recruiter is questioning your honesty, which is a lot harder to recover from than a weak bullet.

How to use AI without sounding like it

• Write the facts first, in your own words: what you worked on, what changed, and how you know. Rough notes are fine.
• Ask AI to shorten a bullet or point out vague phrases. If you ask it to "make this sound more impressive", it will add claims you can't back up.
• Read every bullet out loud. If you wouldn't say it that way to a colleague, rewrite it.
• If you can't say where a number came from, remove the number.

People in the report said they want to show employers more than an application allows, and 52% want to show they can learn and adapt. A bullet about how you noticed a problem and fixed it shows that better than a percentage nobody can check.

If you'd like a second opinion on your bullets, Resumize flags vague and unsupported claims and shows exactly which line each comment refers to. Have you changed how you write applications because of AI? I'm curious what people are cutting.