The rapid expansion of the Internet of Things (IoT) has introduced a complex and interconnected ecosystem that, while enabling intelligent automation, also exposes devices to a broad spectrum of cybersecurity threats. Traditional centralized learning models often rely on aggregating sensitive data in a single repository, increasing the risk of data breaches and privacy violations. This growing concern underscores the need for decentralized approaches that can ensure both effective threat detection and data confidentiality. In this context, the present research introduces a federated learning-based framework designed to achieve privacy-preserving cybersecurity across distributed IoT environments. The proposed system enables IoT devices to collaboratively train a global model without exchanging raw data, thus maintaining individual privacy while enhancing the overall accuracy of cyber threat identification. The study evaluates the effectiveness of this federated approach through performance metrics such as detection accuracy, communication efficiency, and privacy preservation. Results demonstrate that the framework not only mitigates the risks of centralized data exposure but also improves the robustness and adaptability of IoT networks against evolving cyberattacks. The findings highlight that federated intelligence offers a sustainable path toward secure, scalable, and privacy-aware IoT ecosystems, setting a foundation for future advancements in distributed machine learning for cybersecurity.

Keywords: Federated Learning, IoT Security, Privacy Preservation, Cyber Threat Detection, Edge Computing