The recent proliferation of data-driven applications prompted the widespread adoption of federated learning (FL), as it facilitates the collaborative training of deep learning models across various organizations or edge devices without sharing sensitive data with the central server. In FL, participating clients only share their model updates with the central server instead of their raw data. Despite this fundamental benefit, the vanilla implementation of FL does not entirely abide by the current data privacy laws, i.e., the General Data Protection Regulation (GDPR), Health Insurance Portability and Accountability Act (HIPAA), and European Network and Information Security Agency (ENISA), which restrict sharing Personal Identifiable Information (PII). This is because, in vanilla FL, the privacy of local models could be compromised, as a malicious cloud server might try to reverse engineer the submitted model updates to get the PII of the participating clients. In addition, the FL ecosystems are susceptible to client-side attacks, such as model poisoning, inference, and Sybil attacks. This chapter aims to highlight these privacy and security challenges associated with FL. We provide a thorough review of various Privacy-Enhancing Technologies (PETs) proposed to overcome those challenges. In particular, we present a holistic approach to classifying PETs and describe some of the well-known techniques, including Differential Privacy (DP), Homomorphic Encryption (HE), Secure Multiparty computation (SMPC), Zero-Knowledge Proof (ZKP), Private Set Intersection (PSI), Trusted Execution Environment (TEE), Synthetic data (SD), and Distributed Ledger Technologies (DLTs). Finally, we discuss some future directions for enhancing FL privacy and trust by leveraging PETs.