Since the first use of bromide in the 1850s, antiseizure medications have been the first-line treatment for epilepsy. The current paradigm of drug selection largely relies on a “trial and error” approach. Patients may spend years trialing multiple different drugs to try to bring their seizures under control. If seizures continue despite adequate trials of two tolerated and appropriately chosen antiseizure medications, patients are recognized to have drug-resistant epilepsy that carries a substantially increased risk of complications including sudden death. Personalized medicine is an emerging concept in epilepsy that aims to tailor treatments based on an individual’s characteristics. However, the vast amount of multimodal data involved in formulating an optimal epilepsy management strategy for an individual patient, including clinical, genetic, radiological, and electrophysiological data, is beyond the capacity of human or traditional statistical approaches. Artificial intelligence, particularly machine learning, could be a valuable tool to enable personalized epilepsy management. In this chapter, we will focus on two specific aspects when exploring the role of artificial intelligence in pharmacological therapy for epilepsy: selection of antiseizure medications and early prediction of drug-resistant epilepsy. We will review the major studies in these areas, aiming to provide insights into the recent progress in technologies that may herald a paradigm shift in pharmacological treatment of epilepsy.