Artificial Intelligence set to prevent a native German species from becoming extinct in the north
In a bold step towards preserving cultural heritage, a young man from the island of Heligoland, Jakob Martens, has launched the language project Halunder.ai. The aim is to save the Heligoland Frisian language, also known as Halunder, from extinction. Lena Dettmer, a Heligoland native and board member of the Lange Anna Foundation, is pleased with Martens' innovative approach.
The Lange Anna Foundation, committed to the protection of nature and culture on Heligoland for years, plans further projects to save Halunder. These include a learning version for schools and language enthusiasts, and a spoken version to preserve Halunder's sound.
The AI used in Halunder.ai was trained with over one million words, including 40,000 unique words and around 130,000 sentence pairs. AI technology can be used to digitize language resources, provide real-time translation, and support community-led language revitalization efforts.
Specifically, voice recognition and correction apps can help younger generations practice endangered languages interactively. AI-powered translation and dubbing enable real-time learning tools, and platforms like Google’s Woolaroo leverage AI to teach vocabulary in endangered languages via image recognition. Collaborative software projects provide infrastructure that AI can utilize to expand language resources and connect linguistic data globally.
Indigenous researchers and technologists are developing culturally respectful AI tools tailored to the unique needs of individual language communities, reinforcing language sovereignty and identity. For Halunder, similar AI approaches could involve digitizing spoken recordings of fluent speakers, training AI models on the language's phonetics and grammar, and developing interactive educational apps that encourage daily use among younger community members.
Lena Dettmer believes that Martens' work creates a space for future generations to read and understand old Halunder documents. Jakob Martens, on the other hand, believes that while technological solutions alone cannot save a language from extinction, they can make it accessible to future generations who wish to reclaim their cultural heritage. The Lange Anna Foundation actively accompanies and supports the first development phase of Halunder.ai.
The online translator for Halunder, resulting from this training, is data protection-compliant and user-friendly. These training materials were sourced from over 60 years of authentic language sources and reviewed by native speakers. AI models will be trained for these projects in collaboration with Heligoland speakers.
It's worth noting that the Weekly MOPO, a new newspaper with topics like a dream house at a bargain price, a dispute over church asylum, the district of Nord between urban flair and nature, a large puzzle supplement, 20 pages of sports, and 20 pages of Plan7 featuring the sludge metal band The Melvins at the Markthalle, is not directly related to the Halunder.ai project.
In conclusion, the use of AI technology in language revitalization efforts offers a promising future for endangered languages like Halunder. By creating AI-powered tools such as voice recognition, natural language processing, translation models, and interactive learning apps, we can help these languages become practical everyday tools, enhance resource accessibility, and bridge generational gaps where fluent speakers are scarce.
Jakob Martens, fueled by his passion for preserving the Heligoland Frisian language, plans to utilize AI technology to create an interactive learning version of Halunder.ai for schools and language enthusiasts, as well as a spoken version to preserve Halunder's sound, mirroring the efforts of the Lange Anna Foundation. Lena Dettmer, viewing these technological advancements as a means of making Halunder accessible to future generations, envisions these tools providing a way for young community members to read and understand old Halunder documents.
