From ChatGPT to ThreatGPT: Impact of Generative AI in Cybersecurity and Privacy
The evolution of Generative AI (GenAI) models in 2022 has transformed the digital landscape. Models like ChatGPT and Google Bard have grown in complexity, raising cybersecurity concerns. Recent instances have showcased GenAI's use in defensive and offensive cybersecurity, necessitating examination of social, ethical, and privacy implications. This research paper delves into GenAI's limitations, challenges, risks, and opportunities in cybersecurity and privacy. It unveils vulnerabilities in ChatGPT that can be exploited for malicious purposes, such as bypassing ethical constraints. The paper presents successful attacks like Jailbreaks, reverse psychology, and prompt injection on ChatGPT. It investigates how cyber offenders employ GenAI for social engineering, phishing, automated hacking, attack payload generation, and malware creation. Additionally, the paper explores defense techniques utilizing GenAI, such as cyber defense automation, threat intelligence, secure code generation, and attack detection. It also addresses the social, legal, and ethical implications of ChatGPT. Ultimately, the paper underscores the challenges and future directions needed to ensure the secure, safe, trustworthy, and ethical use of GenAI in cybersecurity.