The Taiwan Shinminpo, the only daily newspaper founded and operated by Taiwanese during the Japanese colonial period, was a vital medium through which intellectual youth absorbed modern newspaper practices and culture. It served as a crucial channel for disseminating knowledge, promoting social and cultural movements, and fostering political awakening and cultural enlightenment among the Taiwanese people. First launched in 1923 and approved for daily circulation in 1932, the newspaper endured the forced abolition of its Chinese-language section in 1937—after which it was published entirely in Japanese—yet continued circulation until March 1944.

The Archives of the Institute of Taiwan History, Academia Sinica, have exclusively discovered and completed the digital archiving of historical materials from the Taiwan Shinminpo, including the construction of a comprehensive news headline database. Leveraging AI tools such as Google, ChatGPT, and NotebookLM, we have collaboratively undertaken full-text recognition and Chinese translation of selected columns, including editorials, overseas travelogues, and local feature reports. This AI-processed collection—comprising over 18 million characters of Japanese headlines with corresponding Chinese translations, along with full texts of specific columns—has been fully integrated into the “Taiwan Collectanea Online,” a publicly accessible full-text database.

Given the Taiwan Shinminpo’s uniquely Taiwanese perspective, focus on local concerns, and coverage of diverse socio-cultural issues, it represents an invaluable historical resource for studying Taiwanese movements toward self-governance and observing international conflicts during World War II. Its irreplaceable historical value warrants further digital enhancement, expanded full-text coverage, and broader Chinese translation to support academic research and cultural outreach.

The project is scheduled to be carried out over a one-year period and will initially focus on the following tasks:

1. AI-based text recognition of the full content of morning edition front pages, second pages (editorials), and literary supplement pages, covering approximately 5,000 newspaper pages. This process is expected to generate around 100,000 text entries, totaling approximately 35 million characters of prewar Japanese-language materials.

2. AI-assisted translation and interpretive research on selected newspaper columns, including overseas travelogues and serialized novels.

3. The development of a digital humanities knowledge base grounded in Taiwan’s historical context and archival source materials.