In security, algorithms have become prevalent and used by institutions such as intelligence agencies, police, and courts. These technologies, including facial recognition software, are employed in various security and surveillance practices worldwide. This widespread use raises questions about algorithms epistemic authority and credibility, particularly in producing (in)security practices and contesting evidence within the criminal justice system. In this context, this thesis explores the complex entanglements of the practices of security and legal professionals and algorithms, emphasizing how these digital technologies materialize, stabilize, and circulate in diverse practices even amid errors and contestations. First, the thesis examines the implications of algorithmic reason, addressing how these technologies simultaneously promise efficiency and objectivity while repeatedly getting it wrong. It then explores how algorithms shape perceptions, identify targets, and influence security actions, focusing primarily on biometric data and facial recognition algorithms, such as the use of Clearview AI in the United States. By analyzing these systems, the research aims to understand how algorithms create and legitimize better justice/security imaginaries and their broader social and political consequences. The thesis is located within Critical Security Studies, Science and Technology Studies, and feminist critiques of technoscience, crossing different fields to understand the operative characteristic of algorithmic reason in international politics. Finally, the research demonstrates how algorithms create conditions of possibility for security and justice practices, organizing a multitude of elements and producing an order that impacts these fields and highlights the importance of understanding the political force of the discourses surrounding algorithms and their role in reformulating the conditions of possibility for thinking and doing security.