This document describes the mechanisms deployed in MARVEL to preserve data privacy in the Edge-to-Fog-to-Cloud (E2F2C) computing continuum. This public document describes the efforts carried out under Task 3.1 for the development of “AI-based methods for audio-visual data privacy”. The main goal of the task is to work on methods and algorithms capable to run at the edge and ensure protection of personal data in audio-visual recordings. Although the task focused on solutions for audio and video anonymisation capable of preserving the scene information while removing any information about the identity of the citizens being recorded, privacy preservation in the E2F2C computing continuum is ensured in the MARVEL project in multiple ways. Therefore, even though the core part of the document will be on video and audio anonymisation, we will also provide short descriptions of encryptions and cybersecurity methods (T3.4), approaches for porting the AI processing at the edge (T3.3 and T3.5), and security and authentication mechanisms implemented in the MARVEL Data Corpus in WP2. In this way, we aim at highlighting the connections of those approaches to the ones developed in T3.1 towards increased privacy preservation in MARVEL. To better contextualise the motivations and the challenges behind the adopted solutions, the document also revises the pilot-specific privacy issues that the project has to consider.