Alzheimer’s Disease (AD) is the seventh leading cause of death and the main cause of dementia, yet there still lacks a comprehensive protocol for detecting AD early. Current diagnostic limitations and the availability of large T1-weighted MRI datasets has caused a surge in applying AI to AD diagnosis. In this article, Basata-Torres et al. conduct review research focusing on AI tools utilizing T1-weighted MRI in AD diagnosis, specifically analyzing l the effectivity of models, the superiority of algorithms, pre-processing techniques, and their application of current AI diagnostics in routine clinical practice for AD. Selected through a strict eligibility criteria and following the PRISMA protocol, ninety review articles were selected (n= 90), and their performance metrics were calculated based on their algorithm classification (CNN, ANN, or ML). The studies using CNNs (n=53) showed the highest level of performance, with an average accuracy of 85.93% (median: 87.70%; range: 51.8%-100%). Studies using ANNs (n=9) had a lower mean accuracy of 83.35% (median: 86.67%; range: 57.6% - 99.38%), while studies that used traditional ML (n=28) had an accuracy averaged at 84.22% (median: 87.75%; range: 33% - 99.10%). Overall, similar trends in other performance metrics such as specificity, sensitivity, and AUC showed that CNNs tended to perform better on raw images with a few exceptions. However, there were statistically no differences between the algorithm groups. All studies showed difficulty in clinically differentiating early stages of AD, preprocessing protocols varied greatly among models, and most models were trained on the ADNI dataset, raising concerns regarding generalizability. Further research is required to develop higher-quality AI tools for AD diagnosis through T1-weighted MRI.