A Survey on Privacy-Enhancing Techniques in the Era of Artificial Intelligence

In the era of Big Data and Artificial Intelligence (AI), the unprecedented scale and complexity of data collection, processing, and analysis pose significant privacy challenges. This paper presents a survey, providing a comprehensive and focused overview of privacy-enhancing technologies (PETs) designed to mitigate these challenges and ensure the protection of sensitive information. We explore cryptographic and non-cryptographic techniques including differential privacy (DP), homomorphic encryption (HE), secure multi-party computation (SMPC), and federated learning (FL). Each of these techniques is examined in terms of its basic principles and advantages. Also, some key challenges and ethical issues for implementing PETs are briefly discussed. Finally, we conclude the survey by denoting our future research directions in the field.