Large Language Model for Automated Content Analysis of Online News on Negative Human Tiger Interactions in Indonesia

Newspaper archives provide invaluable cultural, social, and historical information. Human-wildlife interaction is one particular information available on newspaper. This research evaluates the capability of Large Language Models in analyzing Indonesian language online news content related to human-tiger interactions on Sumatra Island. The study collected 1,726 news articles from Google News Archive between 2022 and 2023, with 1,159 articles specifically covering negative human-tiger interactions. Content analysis revealed six interconnected themes: incidence of negative interactions, socio-cultural aspects of communities, authorities' response and mitigation efforts, tiger behavior and health, habitat and population conditions, and illegal activities with law enforcement. The findings highlight two critical threats to tiger survival: extensive habitat destruction and widespread use of pig snares. However, the study also found that local communities' reverence for tigers as respected animals provides a foundation for building tolerance. While authorities have made efforts in conflict response, mitigation measures, and law enforcement, these remain insufficient to address ongoing habitat destruction. This research contributes to the application of Large Language Models for ecological studies of non-English texts.