This research attempts to deploy a Task allocation system using Machine Learning and IoT data streams, and to implement a smart task allocation system for a manufacturing environment, ensuring fair and efficient task distribution among robotic and human resources, in an interest to protect jobs while ensuring efficiency. Three ML Models are deployed for testing using XGBoost, CatBoost, and LightGBM trained on a combined SECOM and industrial robotics dataset, and is fine-tuned using Optuna for hyperparameter optimization and finally chosen based on accuracy, fairness, and production efficiency over a 7-day simulation period. The daily metrics, such as daily efficiency scores, accuracy and fairness of task distribution are monitored. The final aim of this study is to be able to derive results such that this can be applied in a Responsible Industrial AI system later on, which focuses on efficiency without jeopardizing human jobs.