Advancements in artificial intelligence and machine learning for poultry farming: Applications, challenges, and future prospects

• Integration of AI and IoT enhances poultry health management and operational efficiency • YOLOv8 and CNN models improve early disease detection in poultry farming • AI-driven behavioral analysis supports better welfare monitoring and disease prevention • Real-time IoT sensor data optimizes poultry environmental and health management • Edge-AI solutions reduce latency and enable on-site poultry farm monitoring Rapid advancements in Artificial Intelligence (AI), Machine Learning (ML), and the Internet of Things (IoT) have transformed modern poultry farming by enabling precision management and real-time decision-making. This systematic literature review investigates how integrating AI/ML with sensor-based IoT technologies can address key challenges in poultry production, including disease outbreaks, welfare monitoring, and environmental management. The findings reveal that Convolutional Neural Networks (CNN), especially YOLOv8, offer superior performance in visual-based poultry health detection, achieving over 90% accuracy for conditions like bumblefoot and woody breast. IoT-enabled sensors continuously track temperature, humidity, and flock activity, allowing automated systems to adjust ventilation, lighting, and feed distribution. Edge-AI solutions further reduce latency and dependence on cloud computing, making on-site monitoring feasible in low-connectivity environments. This review introduces a structured taxonomy of AI/ML applications grouped into five domains: disease detection, behaviour analysis, environmental control, automation, and productivity optimization. A comparative table of state-of-the-art techniques and their respective performance metrics is provided to highlight the strengths and limitations of each approach. Challenges such as limited infrastructure, high initial costs, and digital illiteracy among smallholders hinder broader adoption. Research gaps are identified in dataset generalizability, multi-modal sensor fusion, and the lack of scalable real-world trials. Future directions include developing lightweight, energy-efficient models, enhancing user-friendly interfaces, and promoting policy frameworks that support ethical and inclusive AI adoption. By critically evaluating current methodologies and outlining actionable strategies, this review contributes to advancing sustainable, efficient, and welfare-oriented poultry farming systems through smart technologies.