The accuracy of battery charge readings in electric vehicles (EVs) is crucial for passenger safety, as incorrect measurements can lead to dangerous situations like overcharging and battery fires. While traditional systems rely on rigid calculations, researchers are exploring the integration of artificial intelligence (AI) to improve battery estimation. However, AI's black box nature raises concerns about reliability. Tests have shown that AI can produce significantly inaccurate outputs when fed corrupted data, highlighting the need for robust safety measures. A proposed solution is a 'safety cage' that monitors AI performance and ensures safety during operation.