Deepfake: A Study on Knowledge of Media Practitioners in Cotabato Province, Philippines

Purpose of the study: Deepfake technology uses an Artificial Intelligence algorithm to convincingly manipulate images, videos, and audio thereby replacing individual's likeness with that of another. The primary objective of the study is to assess the level of knowledge regarding deepfakes of media practitioners in Cotabato Province. Additionally, it examines the relationship between the practitioners’ socio-demographic characteristics, such as their years of experience in the media field and the undergraduate academic program taken, and their level of knowledge on deepfake. Methodology: The study employed a descriptive-correlational research design to examine the relationship between selected variables. A cluster random sampling technique was used to determine the sample. The second district of Cotabato Province comprises 15 radio stations, from which 9 stations were randomly selected as representative clusters. A total of 25 media practitioners participated in the study, including news writers, DJs, reporters, and broadcasters. Technicians were intentionally excluded from the sample, as the focus was on individuals directly involved in the production and dissemination of news and information. Main Findings: Findings revealed that media practitioners demonstrated good knowledge of deepfake content, its creation, and the software commonly used for generating deepfakes. However, their knowledge was limited when it came to deepfake detection and the software tools available for identifying such manipulated content. Furthermore, a significant relationship was found between knowledge of deepfake content and the years of experience as a media practitioner. In contrast, no significant correlation was observed between years of experience and knowledge of deepfake creation, detection, or the corresponding software used for either process. Novelty/Originality of this study: Years of experience in media practice correlate positively with deepfake content knowledge, but not with knowledge of detection or creation tools, suggesting that experiential exposure does not necessarily equate to technical proficiency.