Systematic Literature Review on Technology-Based Fact Verification

Misinformation can lead to various adverse outcomes, including financial losses, misunderstandings, and even psychological issues, which can be mitigated through news fact verification. This research addresses a significant gap in the literature by conducting a systematic and comprehensive review of fact verification technologies in Indonesia. The literature review method in this study utilizes the Population, Intervention, Comparison, Outcomes, and Context (PICOC) framework to identify and evaluate research related to fact verification. We found that efforts to develop fact verification datasets in Indonesia remain scarce, with most existing datasets being in English. One study involved an Indonesian language dataset, namely XFEVER, which uses a translated Indonesian dataset for testing but was not included in the training process. The essential components of a fact verification dataset, such as claim, label, and evidence, must be accurately adapted to fit the context of Indonesian local news. Another finding is that while algorithms like RoBERTa and BERT are frequently used, they require further adaptation for the Indonesian context. This presents an opportunity for future research on training and testing fact verification datasets using multilingual languages, particularly Indonesian, to enhance digital literacy and information discernment skills among the Indonesian public.