尽管大型科技公司的人工智能模型在训练数据中严重缺少少数民族语言，导致这些语言在AI输出中被误代表或完全缺失，但一些社区正在反向利用AI技术创建由本地主导的语言保护平台1。这些创新举措让少数民族社区能够自主构建和管理自己的语言工具，无需依赖外部资金支持1。
巴布亚新几内亚的Vavanagi平台已积累超过80名用户，共贡献超过12,000份英文-Hula语翻译1。Hula语言约有10,000名使用者1，该项目的长老Alu Rigo Ravu Siro对此表示："Just come, join in"1。类似的社区驱动项目还包括东帝汶的Tulun平台和南苏丹的丁卡-英文字典1。其中Tetun语言有超过100万名使用者1，这些平台为面临数字化挑战的语言社区提供了新的可能性。
Charles Darwin University的土著计算机科学家Cat Kutay指出，多个第一民族正在开发自己的语言技术1。肯尼亚和尼日利亚的AI采用率与美国相当1，社区主导的语言保护项目正在成为应对全球AI发展不均衡的一种途径。
Artificial intelligence systems have historically underrepresented minority languages in their training data, leading to gaps in how these languages are processed and translated 1. However, indigenous and minority communities are increasingly leveraging AI tools to create their own language documentation and preservation platforms, bypassing reliance on external funding and corporate solutions 1.
Community-led initiatives are taking shape across the globe. The Vavanagi platform in Papua New Guinea has attracted over 80 users who have collectively contributed more than 12,000 English-Hula translations, supporting the approximately 10,000 speakers of the Hula language 1. Similar efforts include the Tulun platform in East Timor, which serves Tetun speakers numbering over one million, and a Dinka-English dictionary project in South Sudan 1. These platforms empower minority language communities to independently build and manage their own digital language tools 1. Cat Kutay, an indigenous computer scientist at Charles Darwin University, noted that multiple First Nations groups are developing their own language technologies 1. The collaborative spirit driving these efforts was captured by Hula elder Alu Rigo Ravu Siro, who described the initiative simply as: "Just come, join in" 1.
The emergence of these grassroots solutions demonstrates that despite AI's current limitations in minority language representation, affected communities are actively reshaping the landscape of language preservation through their own technological innovation 1.
评论
还没有评论，欢迎留下第一条。