Amongst the many different forms of cancers, the brain cancer is undoubtedly one of the hardest cancers to detect in its early stages due to issues arising from early diagnosis of cancers using medical images. Even though they play an essential role in diagnosing, traditional diagnosis methods tend to be inconsistent, subjective, and lead to many late diagnoses at last. The AI sector has done quite some good in early diagnosis of brain tumors through deep learning in recent years, making the process faster and more accurate by using effective medical image processing. Some of the significant models that can be used include EfficientNetV2S, EfficientNetB7, U-Net, YOLOv5, and efficient deep Convolutional Neural Networks (CNNs) and Vision Transformers (ViTs). All these models use powerful algorithms like transfer learning and object detection amongst others with high efficiency and accuracy. Further, cutting-edge research is devoted to techniques for better generalisation and these include federated learning, where distributed datasets can be used to train a model, and Explainable AI (XAI), which addresses model interpretability for clinical applications. Currently, there is a major interest in using AI-driven innovations as a tool that could revolutionize the diagnosis of brain cancer by providing more precise and reliable tools for improving detection time, more personalized treatment options, and better survival rates as the number of brain cancer cases keeps on increasing around the world.