The pioneering efforts of Bikram Biruli, a 29-year-old researcher, are drawing significant attention for his innovative use of artificial intelligence (AI) to safeguard the Ho language. Hailing from the Adivasi Ho Matkam Sahi village in Odisha's Mayurbhanj district, Biruli's commitment to cultural heritage conservation demonstrates how cutting-edge technology can bolster ancient traditions. His academic journey culminated in September 2025 when he successfully completed his doctoral studies at the Kalinga Institute of Social Sciences (KiSS) in Odisha. This achievement marks a historic milestone, establishing him as the first researcher from the Ho community to employ AI specifically for linguistic preservation.
Biruli's path was rooted in personal experience. He recalls tending cattle and using chalk on a slate to practice his mother tongue, an early immersion that fueled his ambitious project: achieving machine recognition for Ho words comparable to that of English or Hindi. The Ho language belongs to the Austroasiatic family. Data from the 2011 census indicates that approximately 1.4 million people speak Ho, primarily concentrated in the states of Jharkhand and Odisha. Although the language was granted "second official language" status in Jharkhand in 2011, it remains excluded from the Eighth Schedule of the Indian Constitution. This exclusion denies it the highest level of governmental protection. For many members of the community, the language is intrinsically linked to their survival and collective memory.
For his doctoral research, Biruli engineered a sophisticated AI model focusing on three critical linguistic areas. He utilized Automatic Speech Recognition (ASR) technology to transcribe spoken Ho into text, incorporated Entity Recognition to categorize word types, and implemented Part-of-Speech (POS) Tagging to analyze grammatical structures. To train this complex model, approximately 20,000 sentences in the Ho language were processed. Crucially, this training data was recorded using the unique Warang Kshiti script. This indigenous writing system, devised by Lako Bodra, is the preferred choice of native speakers over alternatives based on Devanagari or Latin scripts.
Biruli's work stands as a powerful rebuttal to existing cultural stereotypes, aiming to elevate the language's standing within the digital realm. His immediate plans involve approaching India's Ministry of Tribal Affairs with a proposal to integrate his AI model into official government digital platforms. He is actively seeking to incorporate Ho into applications such as Adi Vaani, which already supports tribal languages like Gondi, Bhili, Mundari, and Santali, and also into Google Translate. The successful implementation of this initiative would set a crucial precedent, enhancing the digital visibility and dignity of tribal languages across the country and opening up new avenues for their transmission to future generations.