The Next Generation of Cyber-Enabled Information Warfare

Malign influence campaigns leveraging cyber capabilities have caused significant political disruption in the United States and elsewhere; but the next generation of campaigns could be considerably more damaging as a result of the widespread use of machine learning.Current methods for successfully waging these campaigns depend on labour-intensive human interaction with targets. The introduction of machine learning, and potentially artificial intelligence (AI), will vastly enhance capabilities for automating the reaching of mass audiences with tailored and plausible content. Consequently, they will render malicious actors even more powerful.Tools for making use of machine learning in information operations are developing at an extraordinarily rapid pace, and are becoming rapidly more available and affordable for a much wider variety of users. Until early 2018 it was assumed that the utilisation of AI methods by cyber criminals was not to be expected soon, because those methods rely on vast datasets, correspondingly vast computational power, or both, and demanded highly specialised skills and knowledge. However, in 2019 these assumptions proved invalid, as datasets and computing power were democratised and freely available tools obviated the need for special skills. It is reasonable to assume that this process will continue, transforming the landscape of deception, disinformation and influence online.This article assesses the state of AI-enhanced cyber and information operations in late 2019 and investigates whether this may represent the beginnings of substantial and dangerous trends over the next decade. Areas to be considered include: social media campaigns using deepfakes; deepfake-enabled CEO fraud; machine-generated political astroturfing; and computers responding to the emotional state of those interacting with them, enabling automated, artificial humanoid disinformation campaigns.