This systematic literature review examines the philosophy of science approaches to user security in distributed devices, such as IoT and Federated Learning. The review was conducted in response to the exponential growth of connected devices and the increasing security threats, including cyberattacks, data breaches, and unauthorized access. As distributed systems become more complex, traditional security approaches, such as cryptography and differential privacy, are often insufficient to address the ethical, philosophical, and contextual challenges that arise in these ecosystems. Distributed devices, especially in IoT and Federated Learning contexts, rely on vast amounts of personal data. This data, often stored or processed in decentralized environments, creates significant risks to user privacy and system integrity. As the number of connected devices grows, security risks multiply, creating challenges in maintaining user trust, privacy, and overall system resilience. Conventional techniques, such as encryption, only focus on technical aspects, often neglecting the deeper philosophical dimensions, such as the nature of knowledge, privacy, and fairness in these systems. These gaps highlight the need for a more nuanced approach that incorporates philosophical perspectives into security frameworks. This study uses a systematic literature review method based on the PICOC (Population, Intervention, Comparison, Outcome, Context) framework to analyze the relevance of epistemology, ontology, and ethics in strengthening system security. By examining the foundational principles of how knowledge is constructed (epistemology), what entities exist in the system (ontology), and the ethical considerations around data and user privacy (ethics), the review provides a comprehensive understanding of how philosophical concepts can be integrated into the design and implementation of security systems in distributed environments. The results reveal that epistemological principles, such as the verification and validation of data sources and models, can significantly improve the reliability and trustworthiness of distributed systems.