ABSTRACT
This chapter considers whether the family of technologies and methods that are commonly thought of as constituting artificial intelligence (AI) could pose a form of threat. AI as a field of study is commonly thought of as being concerned with creating a system that has the capability to learn and reason like a human, but which is not human. For AI to be a threat, targets or victims would need to be affected by the AI negatively, but the harm that they experience does not need to be constrained in any way to cyberspace. The chapter requires the means to identify and manage risks as they emerge, and such risks could apply to all apertures: individual, organisational, national or societal, and global. When considering the threat from AI as a weapon for cyberattack, we do so with reference to the six key characteristics of cyberattacks: targetability, controllability, persistence, effects, covertness and (un)mitigatability.
