Robotic Revival: Can AI be used to revitalize endangered languages?

Written By Malhaar Nair

Cover Art by GrimDreamArt on Deviant Art

Upon waking up every morning, several mobile users look at their phone to see the same notification.“Did you practice your Spanish today?” 

This experience is one I have felt, opening my phone and proceeding to complete my daily language practice. Learning languages has always been an enjoyable pastime for me, one that is now made even more accessible with the rising presence of digital applications catered to this study. AI is evident that the rise of digital teaching is not well discussed, for its simplicity and ease of use allow users to separate their minds from work behind the scenes. However, it is crucial to explore how technology has been utilized in helping individuals learn languages. As the world becomes increasingly interconnected, the role of technology, particularly Artificial Intelligence (AI), is crucial in shaping the way we acquire and master languages In this article, I will delve into the profound impact of AI on language acquisition, exploring how cutting-edge technologies are revolutionizing the way we learn and communicate.

The impending presence of AI in our society is undeniable. Yet, within this inevitability lies a plethora of possibilities to improve and address countless issues across the socio-political spectrum. One pressing concern is the alarming pace of language erasure, as thousands of languages are marked as endangered. In this linguistic crisis, AI emerges as a crucial tool. Its potential unfolds not only in digital infrastructure and data consolidation but also as a transformative force in language preservation, crucial in the field of language endangerment prevention. By providing and amplifying dynamic platforms for learners, AI fosters fluency and contributes to revitalizing countless communities, bridging the gap between technology and cultural connections to languages that risk fading into obscurity.

As of 2022, UNESCO recently stated that there are over 3,000 languages that are classified as “endangered” (United Nations, 2022). With most of them going extinct or “dormant,” only a few in the past century have been revived with great efforts and pain. In a world increasingly focusing on globalism, commerce, and growth, linguists are greatly concerned with the loss of languages as it represents a profound erosion of cultural diversity and a depletion of the collective human intellectual heritage. The extinction of languages implies not only the disappearance of unique forms of expression but also the loss of invaluable traditional knowledge and perspectives.

In the depths of the Southeast Hemisphere, the Maori people and their language in Aotearoa face these very concerns. The struggle to preserve Te Reo Maori is a microcosm of the global battle against language extinction, highlighting the urgent need for innovative solutions. With its ability to facilitate language learning and preservation, AI emerges as a beacon of hope in the efforts to revitalize endangered languages and safeguard the rich tapestry of human linguistic diversity. 

Known as the taonga, or treasure of their culture, Te Reo Maori is the language of the Maori people. Due to the advent of colonialism, policies, and cultural shifts during the past two centuries, the language has been struggling to survive (Reedy, 2000). The Decreasing number of fluent Maori speakers threatens the preservation of this unique language (O Toole, 2020). The only current efforts to revitalize Te Reo Maori include language immersion programs, community initiatives, and educational reforms. Within the Island, organizations like the Maori Language Commission have been working together with the local government to address current problems within the island for the past few decades. With an issue as complex and large as this, the more resources, the better. Thus, integrating AI into these efforts to provide innovative solutions to address the challenges of language learning and retention. For example, AI can help personalize learning for learners depending on what they need to focus on. This is something that has been implemented in many language learning apps on broader platforms (Duolingo, 2023.) Furthermore, utilizing AI can help ease burdens for linguistics and help amplify the work that is being done already. 

It is important to highlight that the number of speakers is not necessarily indicative of the status of a language.(Crystal, 2000). As many indigenous populations are isolated and a large proportion of the population is young speakers, their languages are less likely to die out in the next generation despite their low popularity. Generally, the common thread for any endangered language is the scarcity of speakers in the younger generations. For countless minority communities, this concern is not merely about the numerical count of speakers but also about preserving cultural heritage and identity. The lifeline of a language is not just about the sheer numbers, but also the role of a carrier of unique traditions, values, and ways of thinking (Suina, 2004). For most cases, indigenous languages are the very repositories of indigenous knowledge, reflecting the deep connection between language and the environment, traditional practices, and historical narratives.

So how is a language lost? The key indicator is based on generational transfer. If one generation that is fluent successfully passes the language down to the next, the language transfer could be considered successful (Suina, 2004). Revival of languages is of the utmost importance from a sociocultural perspective. For many communities, language is key to their ahurea, their culture. When a language is lost, the very fabric of the culture is in question. Knowledge that is passed down from generation to generation can be lost. For countless indigenous communities, that loss of knowledge results in impairment of self-identity, self-esteem, and empowerment (Faruk and Rosenbaum, 2022). 

While Maori is not a reo(language) that is critically endangered, it has taken a while to build up a healthy-speaking community. During the 1980s, many activists and organizers started collecting data on the language's health (Reedy, 2000). They conclude that to reintegrate the language into society, seeing was believing: The more people speak the language daily, the more likely the community found a purpose in reclaiming the language. Today, many Maori people have reclaimed their indigenous name with pride, showing that changing community attitudes can allow for the reintroduction of language to occur more naturally (Reedy, 2000). While the Maori have been able to collect data consistently on the status of their Teo, other linguistics communities rarely have that chance. For languages to have a sustainable long-term model, data is the key. Today, agencies and organizations are putting efforts into creating accessible digital archives of the language (Spolsky, 2003). 

Within AI-powered interactive platforms, learners can engage in immersive language experiences. For example, virtual reality (VR) or augmented reality (AR) elements transport users to virtual Maori-speaking environments. This enhances language skills and fosters a deeper cultural understanding, connecting learners with the traditions and stories embedded in Te Reo Maori. Moreover, AI algorithms can personalize learning experiences, adapting to individual proficiency levels and preferences and ensuring a more effective and engaging educational journey (Xie, 2021). The Real-time feedback provided by AI tutors can help learners refine their pronunciation and comprehension, creating a dynamic and interactive language learning environment beyond traditional methods.

AI can provide real-time assessment and feedback on language proficiency, allowing learners to track their progress and improve efficiently. This instant feedback loop is crucial for language acquisition, as it helps learners correct errors and reinforce correct usage. In the context of Te Reo Maori, this feature can be instrumental in refining pronunciation and mastering unique linguistic nuances. Since Te Reo Maori is an oral language, it is key to consider the importance of audio archiving. Digital corpora, large and organized linguistic data collections, are fundamental for linguistic research and analysis (Seyfeydinpur & Nathan, 2011). AI tools can assist in building extensive corpora for Te Reo Maori by processing vast amounts of text and audio data. This corpus linguistics approach helps linguists study language patterns, variations, and changes over time. AI-driven speech recognition and machine translation technologies significantly ease the process of transcribing and translating spoken and written Te Reo Maori. This functionality proves invaluable for converting oral histories, traditional stories, and written texts into digital formats, forming the basis for comprehensive language archives (Bird, 2009).

One “notoriously annoying” but popular language learning platform, Duolingo, has incorporated indigenous languages such as Navajo and Hawaiian into its platform (Anapol, 2018). The whakatūpato(success) of these initiatives, showcases the potential for similar efforts to revitalize Te Reo Maori. The gamified approach of Duolingo, combined with AI-driven adaptive learning, could be adapted to create a dedicated platform for learning Te Reo Maori. Another application, Rosetta Stone, has been involved in language revitalization programs, particularly with indigenous languages. Their approach involves working closely with communities, linguists, and educators to develop language-learning resources(Rosetta Stone, 2023). A similar collaborative effort could be undertaken to create AI-driven resources for Te Reo Maori. 

But where could AI fall behind compared to a human kaiako, or teacher? It all depends on the data. AI makes decisions  based on the data feed it gets. Despite the remarkable capabilities of AI, there are certain areas where it may fall behind compared to a human teacher, and a critical factor in this discrepancy is the nature and quality of the data it relies on(Seyfeddinipur, Nathan, 2011). While AI excels at processing vast amounts of information quickly, its effectiveness hinges on the comprehensiveness and accuracy of the data it is trained on.As stipulated by the World Endangered Language Council, a thriving younger population proficient in the language is paramount for language preservation(Crystal, 2000). In the case of Te Reo Maori, this prerequisite for linguistic vitality is gradually coming to fruition. Presently, a myriad of initiatives are underway, contributing to the rejuvenation of Te Reo Maori across New Zealand. This whakapapa, or resurgence, is not merely coincidental but is a result of deliberate efforts to reintegrate the language into society. Initiatives such as Maori radio broadcasts, language schools, and the renaming of places exemplify the multifaceted approach required for sustained language revival (Spolsky, 2003). By leveraging AI-driven tools for language learning, acquiring fluency in Te Reo Maori can be made more accessible and engaging for the younger population. Interactive language apps, personalized learning modules, and AI-powered tutoring can provide invaluable support to learners, fostering a deeper connection between the youth and their indigenous language. This mahitahi, or synergy, between traditional revitalization efforts and cutting-edge technology, underscores a holistic approach to language revival where AI acts as a catalyst in ensuring the fluency and sustainability of Te Reo Maori for generations to come.

The integration of AI into language learning holds tremendous promise and signifies a crucial juncture in the ongoing haerenga, or journey, to revitalize Te Reo Maori in New Zealand. Through the potent combination of personalized learning, interactive platforms, and cultural sensitivity, AI emerges as a powerful ally in preserving and growing this distinctive language. However, the key to success is fostering collaboration among the Maori community, educators, and AI developers. Together, they can navigate the intricate intersection of technology and cultural heritage, ensuring that the digital evolution enhances rather than dilutes the essence of Te Reo Maori. As we embark on this transformative technological journey, let us remain vigilant in our commitment to preserving and enriching the future of Te Reo Maori, fostering a legacy that seamlessly blends tradition with innovation. There is so much to explore in this, but for now, we must say haere rā: goodbye.

…

Md Omar Faruk, Simon Rosenbaum, The mental health consequences of indigenous language loss, The Lancet Psychiatry, Volume 9, Issue 10, 2022, Page e46, ISSN 2215-0366,https://www.thelancet.com/journals/lanpsy/article/PIIS2215-0366(22)00300-5/fulltext, 

Smith, M. S. (2017). Working with working memory and language. Second Language Research, 33(3), 291–297. 

Suina, J. H. (2004). Native Language Teachers in a Struggle for Language and Cultural Survival. Anthropology & Education Quarterly, 35(3), 281–302. 

Reedy, Tamati. “Te Reo Māori: The Past 20 Years and Looking Forward.” Oceanic Linguistics, vol. 39, no. 1, 2000, pp. 157–69., https://doi.org/10.2307/3623221. Accessed 1 Nov. 2023.

Seyfeddinipur, Mandana, and David Nathan. "Language Documentation in the Age of the Web: Corpus Development and Preservation in the Endangered Languages Archive (ELAR)." Language Documentation & Conservation, vol. 5, 2011, pp. 187-207.

Bird, Steven. "Creating a Comprehensive Language Archive." Language Documentation & Conservation, vol. 3, 2009, pp. 157-175.

Spolsky, Bernard. “Reassessing Māori Regeneration.” Language in Society, vol. 32, no. 4, 2003, pp. 553–78.

O’Toole, Michelle. “Responsibility, Language Movement, and Social Transformation: The Shifting Value of Te Reo for Non-Māori in Aotearoa New Zealand.” Responsibility and Language Practices in Place, edited by Laura Siragusa and Jenanne K. Ferguson, vol. 5, Finnish Literature Society, 2020, pp. 195–212.

Weinhardt, Michael. “Ethical Issues in the Use of Big Data for Social Research.” Historical Social Research / Historische Sozialforschung, vol. 45, no. 3, 2020, pp. 342–68.

Xie, Haoran, et al. “Editorial Note: From Conventional AI to Modern AI in Education: Re-Examining AI and Analytic Techniques for Teaching and Learning.” Educational Technology & Society, vol. 24, no. 3, 2021, pp. 85–88.

Anapol, Avery. “Foreign Language App Duolingo Adds Indigenous Languages to Mark Columbus Day.” The Hill, The Hill, 8 Oct. 2018, thehill.com/policy/technology/410400-foreign-language-app-duolingo-adds-indigenous-languages-to-mark-columbus/. 

Rosetta Stone. “How Rosetta Stone Is Helping Indigenous Communities Reclaim Their Languages.” Rosetta Stone, 9 Oct. 2023, blog.rosettastone.com/how-rosetta-stone-is-helping-indigenous-communities-reclaim-their-languages/. 

Crystal, David. Language Death. Cambridge University Press, 2000. 

“International Decade of Indigenous Languages 2022 – 2032 for Indigenous Peoples.” United Nations, United Nations, 2022, www.un.org/development/desa/indigenouspeoples/indigenous-languages.html. Accessed 1 Nov. 2023. 

Duolingo. “Personalized Learning with the Latest AI - Duolingo’s Approach to Ai.” Youtube, YouTube, 11 Oct. 2023, www.youtube.com/watch?v=-CckdFDptt4.