Bidur Paneru, Anjan Dhungana, Samin Dahal, Lilong Chai
The integration of Artificial Intelligence with sensor networks, computer vision, and predictive analytics enables real-time, data-driven management, improving productivity, health outcomes, and welfare monitoring. An AI-based behavioral, visual, and acoustic monitoring system allows non-invasive, continuous assessment of flock health, enabling earlier interventions and reducing mortality. AI-driven climate control and precision feeding systems optimize resource use, reduce energy consumption, and minimize environmental emissions. Automation and robotics reduce labor dependency, improve biosecurity, and increase consistency in tasks such as egg collection and facility monitoring. Identification of technical, ethical, and adoption barriers provides a roadmap for developing a scalable, explainable, and welfare-oriented Precision Poultry Farming system. The global poultry industry is a critical contributor to food security, providing affordable animal protein to a rapidly growing population. With global meat consumption projected to rise by 14% by 2030, poultry meat is expected to account for a significant share of this increase due to its relatively low environmental impact and production cost (FAO, 2021). A s a result, the poultry industry is under increasing pressure to meet rising demand while keeping high standards of productivity, efficiency, and sustainability. Contemporary poultry production faces diverse challenges, including the need for increased feed efficiency, reduced environmental impact, and enhanced animal welfare (Bist et al., 2024; Choi, 2025). Traditional management approaches, which often depend on subjective human observation, can introduce inconsistencies, and delay the identification of health or welfare issues. Moreover, intensive production systems may heighten stress levels and accelerate disease spread among flocks, underscoring the need for innovative technological solutions that balance productivity with ethical considerations. Historically, livestock farming relied heavily on manual labor and subjective assessments for health monitoring, feed management, and environmental control. Over the past two decades, the integration of sensor technologies and automated systems has laid the foundation for the current era of smart livestock farming. The emergence of artificial intelligence (AI), cloud computing, and Internet of Things (IoT), and edge computing has further enabled real-time monitoring, early disease detection, and precision feeding practices, revolutionizing productivity and animal welfare (Berckmans, 2017). Precision livestock farming technologies address these issues by delivering real-time, data-driven insights that enable prompt and targeted management, ultimately enhancing efficiency, conserving resources, and supporting animal welfare (Schillings et al., 2021; Olejnik et al., 2022). In this context, precision poultry farming (PPF) has emerged as a transformative approach that leverages advanced technologies to optimize the management of poultry operations. Precision poultry farming refers to the application of automated systems and digital technologies to monitor, assess, and manage poultry production processes in real time. Precision poultry farming integrates tools such as sensors, computer vision, robotics, and data analytics to track key parameters, including temperature, humidity, feed and water intake, bird weight, behavior, and health status (Neethirajan and Kemp, 2021). In addition, PPF has emerged as a transformative approach that integrates advanced technologies such as AI, AI-driven machine learning, deep learning (DL), computer vision, IoT, and edge computing, robotics, sensor networks, data analytics, and natural language processing (NLP) to enable real-time, evidence-based decision-making across the production chain. The scope of PPF extends across all stages of the poultry production chain from breeding to broiler houses, egg production and waste management. It enables precise control and optimization of resources, early disease detection, and the improvement of productivity and sustainability metrics. While AI has been applied to a wider range of tasks in poultry farming, this review focuses on key application domains with proven relevance to precision management, animal welfare, and operational efficiency. The primary objective of this review is to provide a comprehensive analysis of AI applications in PPF and to assess their impact on the efficiency, sustainability, and welfare of poultry systems. This review synthesizes current knowledge on the integration of AI technologies in PPF, highlighting their benefits, challenges, and prospects. It emphasizes how AI-driven solutions are transforming poultry management and identifies key areas where innovation can further contribute to the goals of sustainable and welfare-oriented farming. The review addresses the types of AI technologies used, their applications in monitoring and decision-making, and the ethical and regulatory considerations associated with their deployment. Artificial intelligence plays a pivotal role in enhancing the capabilities of precision poultry systems. Artificial intelligence algorithms, particularly those in machine learning (ML) and computer vision, enable the extraction of meaningful patterns from complex datasets generated by PPF tools (Figure 1). For instance, AI can be employed to detect both natural and problematic behavior, predict growth trends, and automate grading and sorting tasks (Neethirajan, 2022). Schematic diagram of the PPF tools applicable in smart poultry farming. Researchers have utilized ML models to detect natural behavior such as dustbathing and perching as well as problematic behavior such as feather pecking and mislaying behavior in Cage-Free (CF) laying hens (Bist et al., 2023c; Subedi et al., 2023; Paneru et al., 2024a, 2024b). By facilitating data-driven decision-making, AI not only improves operational accuracy but also reduces labor costs and enhances the responsiveness of poultry management systems. The integration of AI in poultry farming is a change in thinking from traditional methods to data-driven, automated, and highly efficient systems. Artificial intelligence technologies enable real-time monitoring, intelligent decision-making, and predictive analytics in poultry production systems. Key domains such as ML, DL, Computer Vision, IoT, and edge computing play pivotal roles in modern PPF. These technologies contribute to increased productivity, reduced environmental impact, and enhanced animal welfare (Neethirajan, 2022). Artificial intelligence: Artificial intelligence is defined as the capability of machines to imitate intelligent human behavior, encompassing tasks like learning, reasoning, and problem-solving (Russell and Norvig, 2020). In poultry farming, AI enables automation of complex tasks such as disease diagnosis, behavioral analysis, and performance optimization. Machine Learning: Machine learning is a subset of AI involving algorithms that enable computers to learn from data and improve performance over time without being explicitly programmed (Badillo et al., 2020). For example, ML can be used to predict feed consumption trends or detect anomalies in bird behavior. Deep Learning: Deep learning is a specialized branch of ML that employs neural networks with multiple layers to analyze high-dimensional data such as images and audio. Neethirajan (2022) reviews how DL-based tracking and vision systems are used to assess posture and behavior in poultry farming, and Manikandan and Neethirajan (2025) provide a comprehensive overview of how DL is applied to assess poultry vocalization patterns (including health/disease detection). Computer vision: Computer vision refers to the ability of computers to interpret and process visual information from around the world. In poultry farming, computer vision systems are used to monitor flock movement, detect physical anomalies, and assess crowding or spacing issues (Guo et al., 2020; Cakic et al., 2023; Massari et al., 2022). Robotics: Robotics is the interdisciplinary field that focuses on the design, construction, programming, and intelligent control of physical machines that can sense their environment, make decisions, and perform actions autonomously or semiautonomously, often mimicking or substituting human actions to enhance productivity, efficiency, and safety (Bekey, 2005; Siciliano et al., 2009). Researchers have developed and evaluated a mobile robot system capable of autonomous navigation in poultry houses to assist with labor-intensive management tasks, such as monitoring bird health and removing floor eggs. Field tests demonstrated that the robot could successfully navigate among live chickens with minimal stress to the birds while achieving a 91.57% success rate in automated egg picking (Usher et al., 2017). Internet of Things and edge computing: Internet of Things involves the interconnection of physical devices that collect and exchange data via the internet. In poultry systems, IoT enables the continuous monitoring of parameters like temperature, humidity, feed and water usage, and animal health metrics through sensors and actuators (Wolfert et al., 2017). Sensors placed within poultry houses collect real-time data on environmental and physiological parameters. However, as the volume of data increases, the need for efficient processing and real-time action becomes critical. Edge Computing addresses this by processing data at or near the source of data generation, reducing latency and bandwidth requirements. This is particularly beneficial in remote or rural farm locations with limited cloud access. Edge devices can immediately respond to critical conditions (e.g., ventilation failure or abnormal temperature) without needing to relay data to a central server, thus improving the responsiveness of automated systems (Shi and Dustdar, 2016). Natural Language Processing: Natural language processing is a subfield of AI concerned with the interactions between computers and human language. In livestock/veterinary contexts, NLP has been applied to analyze unstructured textual data such as clinical veterinary reports and free-text health records. It enables automated extraction of insights, improves searchability and summarization of disease trends, and supports decision-making by converting narrative data into structured form (Boguslav et al., 2024; Stimmer et al., 2025). The intensification of poultry farming has raised critical concerns regarding animal health, welfare, and ethical management practices. In this context, the integration of AI tools in PPF has enabled real-time and non-invasive monitoring of bird behavior and health. Adoption of DL-based object detection models, such as YOU ONLY LOOK ONCE (YOLO), has gained popularity among poultry researchers in recent years, and the trend is growing fast. Different versions of YOLO models have been trained and evaluated to detect different behaviors of chickens with high detection precision. For example, researchers have used YOLO models to detect applied and comfort behavior, such as dustbathing (Sozzi et al., 2022; Paneru et al., 2024a), and perching (Paneru et al., 2024b), and problematic behaviors and health issues such as feather pecking (Subedi et al., 2023), piling (Bist et al., 2023a), dead hens (Bist et al., 2023b), mislaying (Bist et al., 2023c), and footpad dermatitis (Bist et al., 2024) in CF laying hens. The applications of YOLO models are not only limited to the CF housing system, but rather it is being used in caged housing, broiler housing, and free-range housing systems to detect various applied and abnormal behaviors of chickens. Technological innovations such as camera-based tracking systems, ML‐based disease prediction, and vocalization analysis, which now play a significant role in enhancing early detection of welfare issues, thereby promoting proactive and precision-based animal care. Computer vision technology is increasingly used to monitor individual birds in a poultry research facility. High-resolution cameras, paired with AI-driven image processing methods, enable analyses of locomotion, spatial distribution, resting versus activity patterns, and interactions among birds. Methods such as object detection, pose estimation, and segmentation are used to differentiate individuals even in moderately dense flocks. These tools provide key behavioral metrics, including activity levels, clustering, and anomalies (e.g., reduced mobility or atypical movement) that often correlate with health or welfare issues. For example, Yang et al. (2024) demonstrated a model that tracks chicken locomotion non-invasively; likewise, Yang et al. (2023) showed the Segment Anything Model’s (SAM) potential in poultry science and laid the foundation for future advancements in chicken segmentation and tracking tasks. Machine learning techniques have become pivotal in predicting disease outbreaks and identifying subclinical signs of illness in poultry populations. These models analyze multivariate data such as environmental conditions, feed and water intake, weight gain, and behavior to detect patterns that precede clinical symptoms (Zhuang and Zhang, 2019). Supervised learning algorithms like decision trees, support vector machines (SVM), and random forests are commonly used for classification tasks, such as distinguishing between healthy and at-risk birds. For instance, real-time data gathered from environmental sensors (e.g., temperature, ammonia levels) and biometric data (e.g., body temperature, movement) can be fed into ML models to predict the likelihood of respiratory infections or heat stress. Importantly, early detection enables prompt interventions, such as adjusting ventilation or administering treatment, thus reducing mortality and improving flock productivity. Furthermore, DL approaches, particularly convolutional neural networks (CNNs), have been employed to analyze image and video data for signs of disease-related behaviors. These models can automatically learn complex features from visual inputs, increasing accuracy in identifying subtle behavioral deviations. A recent study developed a web-based DL pipeline using YOLO11n for disease detection from PCR-verified fecal images (open-source datasets) and EfficientNet-B0 for disease classification, achieving high accuracy (99.12%) and near real-time processing (25.8 ms per image) suitable for farm monitoring. While performance was strong, the dataset’s limited diversity highlights the need for larger, data to improve model and et al., 2025). YOLO object detection model has poultry behavior, and bird A recent review to Computer and in the of research using YOLO models in poultry for various tasks from to as in of of research using models in poultry by research vocalization into the health and of a range of in to environmental or monitoring systems, with AI models, can in and metrics associated with Machine learning algorithms are applied to of chicken to types of and with stress or For instance, increased or may or A study developed a model to automatically detect chicken from achieving over and accuracy while the By techniques to the system potential for real-time welfare monitoring of chicken et al., 2022). in NLP have also been to These data into structured that can be environmental and behavioral metrics, enabling welfare monitoring systems. these in and the need for explainable, AI and sensor integration to enable welfare assessment in poultry systems and 2025). The increasing demand for efficiency, biosecurity, and labor in poultry farming has the integration of automation and robotics into operations. the industry PPF, autonomous systems such as mobile egg collection and are being developed and to assist in a range of from and egg to environmental monitoring and health A study developed a robot with a YOLO vision system and a to automatically detect and floor in a CF housing, achieving over detection accuracy and picking success for both and eggs. of image processing parameters and enabled precise egg extraction and the potential to reduce labor and enhance precision management in a CF system et al., 2021). These technologies not only reduce labor but also increase the and of farm operations. An autonomous mobile robot was developed by et al. and for floor in poultry houses (Figure In with at it successfully only of the or and of the was also to autonomously navigate over in a poultry while and in the of hens. in poultry farming have been developed that can detect floor using image (e.g., or YOLO in or eggs. For instance, et al. developed a robot for free-range that both and with high accuracy under different conditions, and have been used to detect floor even dead in a CF housing system using computer vision but often only detection collection and is For example, Yang et al. (2025) used a with models to detect floor and dead achieving detection in the range of This highlights the future of automation and robotics in the poultry production system. The of floor and autonomous mobile robot are in and a of AI applications in PPF is in in CF laying houses a developed by et al. of AI applications in PPF health, welfare, disease detection, and Artificial intelligence poultry farming over methods, as AI systems can analyze of data in real to and Machine learning models can predict disease outbreaks or performance issues become critical. Automation reduces on manual labor and human animal monitoring allows for early detection of health or behavioral issues, improving bird Artificial intelligence systems improve resource usage, reduce feed energy consumption, and environmental environmental control in poultry houses is critical for improving growth health, and Traditional climate control systems often on or that may not to need or The growing role of AI in poultry farming has climate control methods using sensor networks and data-driven et al., AI-based climate control systems ML models with real-time sensor to environmental parameters such as temperature, humidity, ventilation and ammonia levels, and may also bird behavior or metrics to control actions et al., 2025). A is a or predictive control the system real-time conditions with ML or models to and actuators This approach can bird comfort while reducing energy and reducing on et al., 2022). analytics, a of AI systems, which involves future conditions or behaviors on and real-time In poultry farming, predictive models are employed to optimize and feeding patterns, all of which are critical for performance and ML algorithms analyze data from and sensors to and For example, neural models can predict of high heat and automatically increase or in of achieving high accuracy for and for thus heat stress et al., 2022). AI models can and and on the growth and activity levels of birds. that smart systems improve feed intake, reduce and enhance performance in hens et al., 2020). models assess data on feed intake, body weight, growth and to optimize feeding and This enables precision which reduces and improves feed Furthermore, integration with allows for automated on or et al., 2023; et al., 2017). These the of AI-based technology to automate of environmental conditions, which on the such as housing system, of the growing and this technology has a and potential in the and of these technologies as While AI and precision technologies transformative potential in poultry farming, their a range of These in data and ethical and animal welfare in real-time and barriers to adoption and and these issues is for the sustainable and of AI in global poultry systems. The and are as and datasets are for AI is a of datasets in PPF. For example, (2025) highlights that datasets are relatively datasets with metrics or In addition, smart farming identifies issues such as sensor and of and as to data across (Wolfert et al., 2017). The also showed that data and concerns further data and (Wolfert et al., 2017). Moreover, real-time applications depend on data from sensors temperature, humidity, and feed intake, all of which and data model integration across (Wolfert et al., 2017). concerns and animal AI technologies are often as tools to improve animal welfare, their application ethical For instance, on automated systems could in animal where farm become from interactions with reducing and subtle signs of not by sensors et al., systems may also issues and regarding intensive monitoring in animal Furthermore, ethical the of behavioral tools (e.g., automated or that animal behavior for production While these interventions may increase productivity, be to not the natural behaviors or and in real-time applications AI systems in poultry both and particularly in rural or networks, or cloud for continuous data and remote control (Neethirajan, 2020). While edge computing has been as a to reduce on cloud and edge devices at can be and Moreover, and model significant Deep learning systems, often as may but into their reasoning, which and adoption in critical decision (e.g., disease detection, ventilation and AI models to across environmental conditions, different bird or housing systems, that or is often to accuracy and precision technologies for poultry production and welfare, knowledge be to and research on the and solutions that are and and A concerns the of and datasets for PPF. datasets are and under highly conditions, model on the of and datasets behavior, welfare, environmental parameters, and production metrics. among model the of and is without data is also These are to AI models that can across rather only in research AI and While AI tools are often as improving welfare, is limited on how continuous monitoring, automated decision and behavioral birds and farm over time. research automation reduces or abnormal behavior, and health is a need to to this regarding and behavioral monitoring. of welfare tools that to and on and technological over the past adoption of these technologies research is on analysis across different farm production systems, and even production adoption that technological into Precision Poultry Farming a transformative in how poultry are By AI, computer vision, ML, DL, IoT, edge computing, robotics, and data analytics, PPF enables real-time, evidence-based decision-making that enhances productivity, sustainability, and animal While such as data data model across metrics, ethical and research and support can address these research The integration of AI with technologies and to make poultry farming and in the Paneru is a of Poultry at the of is on enhancing animal welfare and applied behaviors of Cage-Free (CF) laying hens using a data-driven machine vision approach while of the management such as in CF Paneru is of (including and Paneru a of at the Poultry in for a Paneru also a from the of is a in the of Poultry at the of research focuses on poultry welfare monitoring through computer vision and deep learning with applications in automated behavior health and precision livestock farming. has as and to and multiple on poultry monitoring and has research at for artificial intelligence into poultry production systems. has a in the Internet of Things and farm to scalable, solutions for poultry a animal and data analytics to enhance the sustainability and welfare of modern poultry farming. is a of Poultry at the of With a veterinary from the and both clinical and research to research current research focuses on methods to detect and behavioral and welfare of hens in systems, using innovative computer vision and precision livestock farming is of (including and is in the of Poultry at the of is a of the for Precision research and animal precision poultry farming, and poultry health and is of (including and and on as the of Precision Poultry Farming and two poultry at of Precision Farming research and have been with and including and This was for by the of The in this are those of the and not the or of the The study was by and the of for Precision of The of that could have the of this Paneru and and and