Candida species are significant opportunistic fungal pathogens which cause superficial, mucosal and life-threatening invasive infections. Given the potential variation in antifungal susceptibility and drug resistance of non-albicans Candida species, it is imperative to be ableto identify the species at the earliest so that appropriate antifungal treatment can be administered. While all of the above methods are vital to a diagnosis, they can be limited by time, sensitivity of the test, need for trained staff, and lack of availability in low-resource areas. In this context, the possibilities of Artificial Intelligence (AI), machine learning (ML), deep learning (DL) and computer vision for the rapid and automated detection/ identification of Candida are new. AI can be used to process microscopic images, fungal colony morphology, PCR amplification patterns, MALDI-TOF mass spectrometry spectra, outputs from biosensors and antifungal susceptibility data, which may decrease inter-observer variability and enhance the efficiency of the diagnosis. These technologies have the potential to assist in species identification, prediction of resistance, optimizing laboratory workflows, and early epidemiological surveillance. This review explores the new use of AI on the most popular microbiological platforms for the diagnosis of Candida, including the advantages, drawbacks, and clinical implications. It also covers future use of predictive analytics, antifungal susceptibility prediction, multimodal AI, explainable AI, secure data-sharing models, and clinical decision support to enhance the clinical use and accuracy of Candida diagnostics, as well as speed up its delivery.

KEYWORDS: Artificial Intelligence, Candida species, Machine Learning, Deep Learning, PCR, Microscopy, Culture, MALDI-TOF MS, Clinical Mycology.