Cybersecurity Risks: When AI Becomes a Tool For Evil details the multifaceted ways in which AI can be used to wreak havoc on geopolitical structures, social infrastructure, and more. This article mainly focuses on the broad-scale use cases of AI, especially in terms of new and emerging technologies such as GPT-4.
The author outlines the usage of AI to engage in typical adversarial framework such as “election tampering, phishing schemes, and sophisticated man-in-the-middle attacks” —- however, there are even more unique, pressing components that are specifically related to the new capabilities found within these particular types of technologies.
One quite evident example is the usage of AI to produce deep fakes, which are hyper realistic photos, video, images, and sounds that resemble that of reality, to an indistinguishable degree. These can be weaponized in many ways, especially in terms of sowing sociopolitical strife and manufacturing fake information and news.
Another critical component of the usage of AI in this context is advancements in LLMs being utilities to rapidly synthesize complex information and immediately find gaps in codebases and software surrounding critical infrastructure. Finally, those in the space are now using AI to fight AI in the ever-escalating space of the AI Arms Race.
