Rapid advancements in digital technology have significantly impaced many aspects of society, including education. Without a doubt, we are blessed with information and communication technology (ICT). However, it is crucial also to consider its drawbacks. Students are becoming increasingly dependent on digital devices like phones, laptops, and other gadgets in their daily activities. While academic study should always be the top priority, many students today prioritize other online activities, such as gaming, freelancing, and watching movies, over their studies. Our research focuses on the extent to which a student's academic performance can be affected by various input factors using machine learning techniques. In our study, academic performance is categorized into four groups: poor, fair, better, and excellent. This will not only help students take preventative actions based on the predictions from our machine learning models but also enable them to address their academic performance early on. We applied several supervised machine learning models, including K-Nearest Neighbors (KNN), Random Forest (RF), Decision Tree (DT), and Support Vector Machine (SVM). KNN achieved the highest accuracy, reaching 98%. The use of Local Interpretable Model-Agnostic Explanations (LIME) as an Explainable AI (XAI) approach shows the model's interpretability. This study provides students with insights into their current academic performance, labelled as poor, fair, better, or excellent. It allows them to take early action if they believe their use of digital devices is negatively impacting their academic success.