A More Insecure Ecosystem? ChatGPT’s Influence on Cybersecurity

Abstract

With the recent spotlight on ChatGPT, there is now a growing concern about the extent of Artificial Intelligence’s (AI) influence on cybersecurity. Regulatory efforts to manage AI’s risk to individuals, organizations, and society are well underway; however, these are still very much in their infancy. Yet, growing public discourse has called into question the ethics and use of these models; going so far as to suggest pausing AI development. While there remain vocal detractors, AI researchers continue to develop more powerful models independent of regulatory oversight. Under continued social pressures and scrutiny, these organizations implement their own policies in an attempt to stave off external influence. There are many inherent risks of using large language models like ChatGPT (e.g. bias, misinformation, manipulation, and privacy concerns); however, this paper will focus on ChatGPT’s organizational implications on the human in cybersecurity. Are robust AI models like ChatGPT able to alter the current security ecosystem? AI has historically been used to enable technical controls in defensive cybersecurity strategies. Models like ChatGPT are able to blend technical and social constructs, potentially shifting this paradigm. At the same time, these use cases help researchers and regulators better understand the social ramifications. Understanding applications of powerful AI models, even in a niche application like human factors of cybersecurity, can drive regulatory and public discourse.