Research
Prompt Injection in Automated Résumé Screening Reliably Boosts Applicant Rankings
As LLMs increasingly screen and rank job applicants, this study shows that subtle self-promotional text — adding no new qualifications — reliably improves an applicant's ranking under both single- and multi-injection settings, especially when résumé quality is homogeneous and few candidates inject. It quantifies a concrete adversarial incentive baked into algorithmic hiring pipelines. For builders deploying LLM evaluators, it's a direct warning that ranking outputs can be gamed by content, not credentials.
↳ Follow the thread