Synthetic media for multilingual Massive Open Online Courses: accessibility, engagement, and ethical governance

In recent years, synthetic media from deepfake videos has emerged as a new and interesting technology. Synthetic media may refer to cloned voices or multilingual translation models. Avatar-based tutors have emerged as important tools for higher education. Consequently, these technologies are being incorporated into distance learning models and, more recently, Massive Open Online Courses (MOOCs) worldwide, at an increasing rate. Although the term synthetic media is often associated with risks, recent studies suggest that, when used transparently and for pedagogical rather than entertainment objectives, these tools can improve accessibility, personalization, and increase learner engagement. Therefore, this scoping review examines the international literature published between 2020 and 2025 on the use of deepfake and synthetic media tools, primarily in online and blended higher education contexts, and discusses their implications for multilingual MOOCs, where large-scale empirical evidence remains limited. The article also examines ethical and policy issues connected with the adoption of these technologies, and drawing on educational technology research and policy documents, like United Nations Educational, Scientific and Cultural Organization (UNESCO) guidelines and the European Union Artificial Intelligence Act (EU AI Act), the review shows that synthetic avatars and AI-generated videos can reduce production costs and support multilingual learning. However, concerns remain regarding authenticity, privacy, and the changing nature of the teacher–learner relationship. The central contribution of this paper is the proposal of a policy framework focused on transparency, responsible governance, and AI literacy. Consequently, the goal is not to replace human instruction, but to integrate synthetic media in such ways that could strengthen the pedagogical design, safeguard rights and make multilingual MOOCs more engaging rather than an automated system without human participation.