In conversations all over the potential of AI and cyber-threats, we typically speculate when we can assume to see destructive or offensive AI attacks in the wild. Though we have not however noticed conclusive evidence of execution, this report will show that all the tools and open-source exploration essential to aid an AI-augmented assault exist these days. For that reason, we can foresee that AI-pushed cyber-attacks are not decades absent, but a really serious chance in the rapid potential.
This report will doc an conclusion-to-conclusion assault lifecycle, and how each individual phase could leverage features of the AI ‘toolkit’ to increase and streamline the process. Attackers will, of training course, evolve their tools to drive efficiency gains, however these tradecraft advancements are iterative and do not come about all at the moment. In addition, while it is most likely that adversaries these days are now leveraging AI in some ability to increase personal assault phases, this report demonstrates an conclusion-to-conclusion AI-pushed assault purely as a thought experiment