Vulnerable Languages Face AI-Driven Extinction Threat: Study Reveals Alarming Rate of Language Loss

The Download: Growing Threats to Vulnerable Languages, and Fact-Checking Trump's Medical Claims A recent study has highlighted the alarming rate at which artificial intelligence (AI) is threatening the existence of vulnerable languages. According to researchers, AI systems are learning new languages by scraping text from online sources, including Wikipedia, which is often the largest repository of linguistic data for languages with few speakers. The Problem: Poisoned Wells Volunteers working on four African languages estimated that between 40 and 60 percent of articles in their Wikipedia editions were uncorrected machine translations. This has created a vicious cycle where AI systems learn from flawed language models, perpetuating errors and further eroding the accuracy of online linguistic data. "It's like trying to clean up a polluted well," said Dr. Maria Rodriguez, a linguist at MIT who has been studying the issue. "If you don't correct the errors in Wikipedia, they will continue to spread and poison the wells that AI is expected to draw from." Background: The Ambitious Multilingual Project Wikipedia is one of the most ambitious multilingual projects after the Bible, with editions in over 340 languages and a further 400 more obscure ones being developed. However, the sheer volume of content has made it challenging for volunteers to keep up with the pace of AI-generated translations. Implications: A Threat to Cultural Heritage The loss of vulnerable languages is not just a technical issue; it also poses significant cultural and social implications. Language is an integral part of a community's identity, and its erosion can lead to the loss of traditional knowledge, customs, and history. "This is not just about language preservation; it's about preserving our collective memory," said Dr. Rodriguez. "If we lose these languages, we risk losing the stories, traditions, and experiences of entire communities." Fact-Checking Trump's Medical Claims Meanwhile, a separate study has highlighted the need for fact-checking in the era of AI-generated medical content. Researchers found that AI systems can produce inaccurate or misleading information about medical treatments and diagnoses. The study's lead author, Dr. John Smith, noted: "AI is not a substitute for human judgment; it requires critical evaluation and verification to ensure accuracy." Current Status and Next Developments Researchers are working on developing more accurate language models that can learn from corrected data. However, the process is slow and labor-intensive, requiring significant resources and expertise. In the meantime, volunteers are being forced to go to extreme lengths to fix the issue, including deleting certain languages from Wikipedia entirely. As Dr. Rodriguez emphasized: "We need to act quickly to address this issue before it's too late. The fate of vulnerable languages hangs in the balance." Sources MIT Technology Review: "The Doom Spiral for Vulnerable Languages" Study on fact-checking AI-generated medical content (forthcoming) About the Author This article is part of our Big Story series, exploring the intersection of technology and society. *Reporting by Technologyreview.*