Targeted Threats, Persistent Practices: Understanding the Evolution of Digital Repression in Autocracies and Democracies

At the dawn of the Internet age, cyber-optimistic interpretations of digitalisation saw Information and Communication Technologies (ICTs) as liberation technologies for collective mobilisation. Paralleling this, initial views on the advances of Artificial Intelligence (AI) posited it as a promising instrument for societal progress. However, like ICTs, initially empowered as soft weapons of democracy, can devolve into instruments of control when wielded by the oppressors, AI's transformative capabilities exacerbate the constant decline of Internet freedom. To advance scholarship on these countervailing forces, we investigate the dynamic landscape of "digital repression." "Repression" is a broad concept that historically includes actions or policies that are meant to, or do, raise the costs of activism. "Digital repression" is defined as those actions directed at a particular target to increase the costs for online contentious activity and the use of social media to raise the costs for social movement activity, wherever that contestation occurs (Earl, 2022). The "basic" digital repression toolkit includes surveillance (e.g. digital surveillance and spyware), censorship (e.g. shadowbanning and filtering), social manipulation (e.g. disinformation and misinformation), cyberattacks (e.g. phishing, malware, DDOS, ATP), connectivity restrictions (e.g. Internet shutdowns and filtering), and targeted arrests of online activists. The use of AI amplifies at least four of the six digital repression techniques we mentioned, namely, surveillance (e.g. biometric technologies), censorship (e.g. automated identification of threatening communications and content removal), social manipulation (e.g. AI-powered social media bots and AI-generated deep fakes) and cyberattacks against targeted individuals and organisations. Research on digital repression often suffers from a solid authoritarian gaze (i.e. digital authoritarianism as the use of digital information technologies by authoritarian power holders for political and social control). Still, in line with cyberspace's borderless and transnational nature, digital repression is not confined to a specific regime type within a defined territory but also emerges in democracies. While state openness plays a significant role in explaining state repression in response to online free speech, democratic uses of digital repression prove that authoritarian practices are pervasive in non-autocracies (Glasius, 2018). With the aim of disabling access to information or silencing voices, digital authoritarian practices are patterns of actions sabotaging democratic accountability to people over whom a configuration of actors exerts a degree of control. Drawing from the Digital Society Project (DSP) dataset and the AI Incident Database (2010-2022), we provide insights on a concerning trend where authoritarians and democracies exploit technological advancements (e.g. ICTs and AI) for purposes contrary to their intended benefits – from heralded instruments of empowerment to transformative forces reshaping digital domains into arenas of repression. This paper questions how the evolution of digital repression impacts online freedom of expression and political activism; we argue that digital repression translates into targeted digital threats that are not isolated incidents but persistent authoritarian practices that deliberately silence voices (Hirschman, 1972). This paper underscores a transformative trajectory of political repression in autocracies and democracies – no longer violent and overt, yet achieving the exact outcome of silencing opponents.