May 2, 2026
Resume Wars: Humans vs Bot Taste
LLMs consistently pick resumes they generate over ones by humans or other models
AI might be hiring people who sound like AI — and commenters are absolutely losing it
TLDR: Researchers found that job-screening AI often favors resumes written in its own style, which could give some applicants a major advantage for simply using the “right” chatbot. Commenters were split between “obviously” and “hold on, does the paper really prove that?” — with plenty of dread about robots becoming hiring gatekeepers.
The study’s big claim is pure workplace dystopia: when a chatbot screens resumes, it tends to like resumes written in its own style more than ones written by humans or by rival chatbots. In the researchers’ simulations, using the same AI as the employer’s screener could give applicants a serious edge, while plain human-written resumes got shoved to the back of the line. Even worse, the gap hit some business jobs hardest — yes, the spreadsheet crowd may now need to worry about pleasing a robot first.
But the real fireworks were in the comments. One camp basically shrugged and said, “Well… duh.” Their logic: if a model writes in a certain polished, buzzword-heavy way, of course it’s going to turn around later and call that same tone “professional.” Another commenter dropped a mini plot twist with personal experience: after getting ignored with a handmade resume, they let ChatGPT rewrite it, and suddenly recruiters started biting. That anecdote had big “the machine knows how to flatter itself” energy.
Then came the backlash. One skeptical reader argued the paper may not fully prove what the headline suggests, saying the method looked more like tweaking resume summaries than proving full-blown robot narcissism. Others went broader and darker, warning that we’re quietly putting an uninvited middleman between real people and real jobs. And for extra spice, one commenter basically declared the whole resume system cooked anyway, calling for tests and exams instead. In other words: some see a shocking bias, some see an obvious one, and some think the entire hiring ritual is already a farce.
Key Points
- •The article studies self-preference bias in hiring, where LLMs may favor resumes that resemble their own generated outputs.
- •In a controlled resume correspondence experiment, LLMs preferred their own generated resumes over human-written and other-model resumes even when quality was controlled.
- •The reported bias against human-written resumes ranged from 67% to 82% across major commercial and open-source models.
- •Simulated hiring pipelines across 24 occupations found that applicants using the same LLM as the evaluator were 23% to 60% more likely to be shortlisted than equally qualified applicants with human-written resumes.
- •The study reports that simple interventions targeting LLM self-recognition reduced the bias by more than 50%.