Digital Ministry Adds 20 Million Tokens of Hakka Language Data to Taiwan Sovereign AI Training Corpus

The 'Taiwan Sovereign AI Training Corpus,' promoted by the Ministry of Digital Affairs, has recently partnered with the Hakka Affairs Council to officially release valuable Hakka language resources, publications, historical and cultural archives, and research reports. This marks the first addition of approximately 20 million tokens of high-quality training data, injecting greater cultural diversity and linguistic vitality into Taiwan's sovereign AI corpus. This enhancement will improve AI's comprehension of Hakka language and Hakka culture, supporting diverse applications such as intelligent Hakka customer service and educational learning tools.

As one of Taiwan's national languages, Hakka carries profound historical and cultural significance. The materials provided by the Hakka Affairs Council are highly representative, primarily covering: poetry collections, books on Hakka customs and language instruction; vocabulary textbooks and test banks for Hakka proficiency certification, spoken language transcripts, cultural archives, and academic and policy research materials on topics such as Hakka language revitalization, policy development, and cultural governance.

The Ministry of Digital Affairs emphasized that these materials not only reflect Taiwan's achievements in preserving Hakka culture but also enhance AI models' ability to understand and generate content in various Hakka dialects, semantics, and cultural contexts. In the future, these capabilities can be widely applied in intelligent Hakka customer service, local cultural knowledge services, and AI-driven public policy innovations.

Since its launch, the 'Taiwan Sovereign AI Training Corpus' has accumulated over 1.5 billion tokens. The Ministry will continue collaborating with government agencies to release high-quality language data, deepen public-private partnerships, and plan to promote private-sector data collection, gradually building a data foundation that represents Taiwan's sovereignty and professional knowledge.