Visakhapatnam: A group of faculty members and researchers at Centurion University of Technology and Management, Vizianagaram, has launched the final phase of community validation for Mahul Vani AI, a multilingual artificial intelligence platform designed for tribal communities. The initiative is led by the Magicterrain Foundation (MTF), a section 8 non-profit working at the intersection of ecology, indigenous knowledge, education and rural livelihoods. Developed over two years through field research under MTF’s in-house innovation stream, MTF Prayog, the AI can speak, listen and write in more than 30 tribal languages, including Saora, Kandha, Rabha, Toto, Mech, Oraon, Munda, Santhal and the Chenchu language spoken in parts of Andhra Pradesh. Many of these languages remain largely unsupported by mainstream AI systems.According to the researchers, the platform has been trained using more than one lakh entries drawn from tribal and local knowledge systems. The aim is to create an AI tool that understands and responds to people in their own language and local context. The researchers describe it as a system through which a tribal grandmother can ask a question and be understood, or a tribal farmer can photograph a diseased crop leaf and receive guidance in his/her own tongue. The platform is intended to serve remote tribal communities living in hilly and forest regions where access to agricultural officers and other public services is often limited.The first phase of trials in the Buxa Tiger Reserve region of West Bengal is scheduled to begin on June 20. The second phase of trials has been completed in the Gumma block of Gajapati district in Odisha, while baseline data collection has been completed in the Nallamala region of Andhra Pradesh. Field trials among Chenchu tribal communities are expected to begin in June.The project has been conceptualised and guided by Dr Atanu Deb, a faculty member at Centurion University and the founding director of MTF. Sambhav Barik serves as the technical head, while Basabendra Sinha Mahapatra, Jonnala Saikrishna Reddy and Saranya Yadav are part of the research team. The community trial team includes Dr Preetha Bhadra, Purnajyoti Khanra, Subhankar Chhaondogi and A. Divesh Babu. Funding and technical support for the initiative are being extended by J C Kumarappa Technology Park & Innovation Centre LLP.Speaking to ToI, Dr Atanu Deb said the AI platform is being developed specifically for tribal communities and is intended to function at the village level. “Our team has already collected baseline survey data from around 5,000 people, and a larger dataset is being prepared as part of the ongoing community validation process. The system is trained using local village data, including information related to natural resources, farming conditions and community knowledge systems, enabling it to provide highly localised support. For example, the platform can function as a ‘crop doctor’ by helping tribal farmers identify agricultural problems and access advice in their own language,” he said.Dr Deb added that, unlike general AI language tools, Mahul Vani AI is trained using village-specific demographic and environmental data, allowing it to provide personalised responses linked to local land conditions, crops and community practices. He said the platform is intended to bridge the service gap faced by remote tribal populations living in difficult terrain.