Medical devices and machines are increasingly utilizing Cyber-Physical Systems (CPS), bringing about vulnerabilities that could easily be exploited through cyberattacks. Closed loop insulin pumps are especially prone to such activities as the connection between the components used, the pump and Continuous Glucose Monitor (CGM) sensor, may pose a threat to patients as it susceptible to cyberattacks. Due to more and more reliance on these systems, the chances of attacks have increased drastically. This paper proposes a method to help detect intrusions and improve the safety and stability of the system using Generative Adversarial Networks (GANs). Its intrusion detection capabilities strengthen and safeguard users from harm due to insulin pump hijacking attacks. The experimental findings indicate that our deep learning approach surpasses currently existing studies in this domain. The experimental findings indicate that our deep learning approach surpasses currently existing studies in this domain. This study makes a valuable contribution to improving cybersecurity in the healthcare sector by providing a novel solution to protect and ensure that security measures are in place.