Egg is an indispensable food choice to meet the ever-growing demand for food globally. However, the egg industry suffers from the serious ethical, ecological and economic issue of culling of day-old male chicks. Predetermination of egg sex before incubation, also known as in-ovo sexing, is considered an ideal solution to this issue. Despite consistent and continued research in this direction, a commercially viable large-scale application is still lacking to date. Artificial Intelligence (AI) remains promising in this area, offering both invasive and non-invasive approaches. This chapter provides an overview of existing AI-backed approaches for in-ovo sexing in hatching eggs while highlighting their challenges, limitations and future prospects. The current in-ovo sexing approaches include classification based on either internal images, such as hyperspectral images, blood vessel images and embryonic vascular images, or external images of morphological features, yielding an accuracy between 83% and 95% on eggs of varying ages. While AI holds potential, its commercial adoption for standardized large-scale applications is still in its early stages. On the other hand, concerns persist about the reliability of AI due to its major difficulties in obtaining consistent and reliable data across varying egg types and imaging conditions. Based on the review, this study recommends an inclusive approach involving both industry and academia to develop an AI-based classifier with the potential to discern egg gender even before day 4 through a non-invasive method. This approach would focus on image-based classification using morphological features, high throughput, and achieve industry-standard accuracy.