VIDEO CLASSIFICATION USING LARGE LANGUAGE MODELS APPROACH FOR FILM CENSORSHIPS IN INDONESIA

The advancement of technology has made the dissemination of videos increasingly massive and widespread, making it very easy for anyone to access them at this time. Not only the dissemination of videos, but even the creation of videos is becoming easier and can be done by anyone. With the development of the times, now the media for video broadcasting is becoming more and more numerous and diverse with various innovations from each platform provider. Because of this, it is also related to the reach of viewing access. Viewing access that was originally public through media such as television (TV) and cinemas, is now rapidly evolving into media convergence that brings integrated viewing closer to the public. Therefore, an effort is needed to utilize technological advancements to help or alleviate the work of an institution that directly faces the impact of media convergence, one of which is the utilization of artificial intelligence (AI) for film censorship. In film censorship, there are many aspects or factors that influence whether a content or scene is deemed suitable for censorship, some of which are pornography and violence factors, of course, based on the rules established by Law Number 33 of 2009 concerning Film and Government Regulation Number 18 of 2014 concerning the Film Censorship Institution (LSF). In the film industry, the use of AI technology has experienced significant development. One important aspect in film production and distribution is the censorship process to ensure that the film complies with age classification guidelines, ethical norms, and applicable legal regulations. The film censorship process, which involves identifying and removing inappropriate content, has long been a time-consuming task involving significant human resources. With the development of Large Language Models (LLMs) that can be used to analyze video, images, sound, and text, there is potential to automate and improve the efficiency of the film censorship process while ensuring that the censorship quality is maintained.