This article analyzes how Artificial Intelligence (AI) and the Internet of Medical Things (IoMT) can improve diabetes management and outcomes in pregnant women through providing continuous monitoring, remote testing, and automated, personalized care. The multi-author, peer-reviewed article consolidates the findings of several studies, including meta-analyses, observational studies, and randomized controlled studies, to cover the plethora of new technologies aimed at improving diabetes care during pregnancy. Various AI model classes are reviewed, including CNNs in image recognition and transformer NLPs in medical decision-making. Several limitations and possible bias sources were noted. Firstly, black box algorithms lack transparency, making it difficult to identify errors and biases. Secondly, limited data availability, high cost, and lack of representative data mean many models have not been thoroughly tested. This has the potential to exacerbate health disparities, targeting populations with less access to care. Furthermore, few have been validated in clinical settings, raising concerns about their accuracy in real-world applications. Finally, data leaking can lead to inaccurate predictions. The primary endpoints vary by technology. One model predicting gestational diabetes had a 70.3% accuracy, and several studies showed reduced hypoglycemia for patients using continuous glucose monitoring. Many studies also show mixed results. However, the direct impact of these technologies on improving clinical outcomes is still unknown. Overall, these technologies have the potential to revolutionize care through remote monitoring and testing, providing real-time data and automatic but individualized treatments. However, cost, data leakage, and transparency challenges must be solved before they can significantly impact clinical care.