Digital Surveillance, Master Key for MENA Autocrats

In the past decade, a number of MENA countries published national AI strategies (Egypt, Jordan, Lebanon, Morocco, Saudi Arabia, Turkey, UAE, Qatar) or announced their intention to do so (Tunisia, Algeria, Bahrain). Some have established national AI councils or similar dedicated institutions to centrally foster AI across sectors (Egypt, Israel, UAE, Qatar, Saudi Arabia), alongside research centres, networks, and centres of excellence for AI development. Contrary to the emerging European risk-based approach to AI, MENA AI strategies overwhelmingly emphasize the opportunities inherent to AI, but barely assess the risks. Although references to international soft law norms such as the OECD principles on AI are found in almost all documents, there is little space given to tangible ethics concerns, let alone a clear roadmap on how to mitigate them. In the contribution, the authors explore how the uptake of AI-enabled technologies has boosted digital surveillance in the Middle East and North Africa. Major advances in artificial intelligence, including machine learning for clustering, speech recognition and -generation, natural language processing, image and video generation, autonomous decision making and intelligent personal assistance, have given impetus for upgrading both mass and targeted surveillance solutions. Digital surveillance provides a master-key for MENA autocrats, as it facilitates identification, targeting and tracking of dissidents and their associates. The contribution goes on to show how, unlike in other AI segments, China is not yet a lead player in the MENA deep packet inspection (DPI) or spyware markets as MENA governments have treaded carefully not to get caught up in a global techno-political competition between China and the US; leaving the MENA digital surveillance market largely to American, Canadian, European, and Israeli companies. MENA law enforcement and security agencies, telecom and internet service providers have been employing both mass and targeted digital surveillance solutions to track not only criminal suspects but also activists and human rights defenders mashed together by diffusion of blurry cybercrime laws. With the global shift to fifth generation (5G) networks and drastic increases in data capacities and speed, digital surveillance will inevitably rely on further automation and AI algorithms. Without policy interventions, the author sustains, the described surveillance patterns are likely to quickly expand as MENA government agencies will combine different AI-assisted surveillance solutions from an ever-growing pool of tools to conduct targeted and mass surveillance both domestically and abroad. In this vein, the authors points out lines of action policymakers need to take to contain this worrisome trend in digital authoritarianism.