Deepfake news: AI-enabled disinformation as a multi-level public policy challenge

The advent of ‘DeepFake' content that is increasingly difficult for humans and machines to distinguish as artificial portends a number of challenges to democratic societies. In order to effectively respond, policymakers must gain understanding of how DeepFake content might manifest. This paper aims to offer necessary context by exploring AI-enabled multimedia disinformation across different levels: (1) as a mass-produced, regular feature of the information environment in democracies and (2) as a highly tailored instrument used in tandem with cyber operations. I explore the impact of DeepFakes on the ability of populations to determine the origination, credibility, quality and freedom of information. Such macro impacts amplify the potential value of DeepFake content employed alongside targeted cyber activities, a combination that even alone offers belligerent actors new opportunities for enhancing attempts at disinformation and coercion. Nevertheless, I ultimately argue that DeepFakes should be thought of more as an evolution than a revolution in disinformation techniques, the real threat of which emerges from the manner in which new abilities to produce even reasonable fidelity fabrications rapidly and at scale combine the multiform shape of the modern digital information environment to make organized influence efforts much more dynamic than has previously been the case.