Impact of Generative AI in Endpoint security

Abstract

Advanced endpoint security is needed to combat sophisticated cyberattacks. Generative AI methods including GANs, VAEs, and autoregressive models are used to improve endpoint security in this paper. Despite training stability issues, GANs produce high-quality synthetic data for malware detection and attack simulation. With stable training, VAEs detect anomalies but provide lower-quality data. Though computationally costly, autoregressive algorithms detect insider threats and network breaches with excellent accuracy in sequential data analysis. Comparative analysis shows model strengths and limitations, guiding endpoint security framework use. Integrating GAN stability, VAE data quality, and autoregressive model optimization with security measures and hybrid models are future research goals.