Published Paper Details:

Artificial Intelligence (AI) is changing the way companies hire people by making it possible to automatically screen resumes, predict who will get the job, and assess behaviour. In addition to making things much more scalable and efficient, these new technologies also bring up significant ethical concerns, particularly the potential for algorithmic bias.This research presents a practical study on identifying and reducing bias in AI-powered recruitment systems.The research investigated how unfair decisions may occur due to bias in training data or decision-making models,especially with respect to sensitive attributes like gender.Using a dataset of resumes ,the system first detects bias and then applies three types of techniques to reduce it:before training (pre-processing),during model learning(in-processing) and after predictions(post-processing).Methods like Reweighing and Adversial Debiasing were used to improve fairness without significantly affecting accuracy.The fairness of the system was measured using metrics such as demographic parity difference and results showed that debiasing techniques can reduce discrimination in predictions. Additionally ,tools like LIME and SHAP were used to explain the model's decisions, helping users understand why certain resumes were favoured. This research supports the development of fair and transparent AI models for hiring.