Adversarial Prompting and Bias Testing in Hiring AI
Hiring AI systems can encode hiring manager biases in ways that look like objective scoring, so testing these tools with adversarial prompts—submitting identical resumes with different names, backgrounds, or histories—reveals whether the system treats all candidates fairly. Understanding where AI hiring systems are vulnerable protects you from being rejected by flawed automation.
HypatiaAdversarial prompting involves deliberately crafting inputs designed to expose weaknesses, biases, or inconsistencies in an AI system, and in hiring contexts it is used to test whether screening tools disadvantage applicants with criminal records or employment gaps.
Reentry advocates and job seekers who understand adversarial prompting can use it to audit the AI tools they rely on, identify where bias may reduce their visibility, and adjust their application materials to perform more equitably across different screening systems.
Ready to work on Adversarial Prompting and Bias Testing in Hiring AI?
Explore related journeys, or bring what you’re working through to Hypatia.