Antimicrobial peptides (AMPs) are promising candidates for the current antimicrobial resistance public health issue due to their multiple mechanisms of action that enable a broad spectrum of activity. However, screening large peptides is labor-intensive and costly, with low reproducibility and additional challenges due to their toxicity, off-target effects, and low in vivo performance. Given the urgency of finding new molecules, strategies based on in silico screening followed by in vitro validations and toxicity evaluations are a promising pipeline for discovering candidates. Specifically, artificial intelligence (AI) enables the designing or discovery of AMPs through sequence-function relationships. AI can be used early in the discovery process by optimizing the selection of candidates to evaluate in vitro or to analyze the physicochemical characteristics of candidates obtained elsewhere. AI can also generate de-novo peptides from a set of desired characteristics. This chapter elucidates the most used methods, exploring their applications and their respective advantages and disadvantages along the AMPs discovery pipeline.