UK-based voice AI tech firm ElevenLabs is deepening its stronghold in India, with new MoUs announced with state and central governments, as well as an expansion to support all 22 languages. Co-founder and CEO Mati Staniszewski along with Karnataka’s Minister for Information Technology and Biotechnology, Priyank Kharge, formalised the firm’s partnership with an MoU signed with the Karnataka Innovation and Technology Society (KITS).

Aside from unlocking commercial opportunities, voice AI use cases can find a lot of scope in public policy, Nihal Chauhan, head of public sector partnerships and policy for APAC at ElevenLabs, said. “Without access to the right language you can miss out on the right information, like from say, beneficiary schemes,” he noted.

The partnership promises to help push voice AI adoption in Kannada across areas including skilling, citizen services, grievance redressal, and responsible AI. Applications will range from voice-enabled interview practice, upskilling in Indian languages, voice restoration for people who have lost the ability to speak, investor assistance and information services, and voice technology for cultural and public service initiatives.

The company has committed to restoring 1 Million Voices under its initiative for those who have permanently lost their voice due to serious ailments like ALS, cerebral palsy, oral cancer, and multiple sclerosis.

Last week, separately, the firm also signed an MoU with the central government to join a consortium with the Coalition of the Willing, while working with the Department of Empowerment of Persons with Disabilities, The Quantum Hub, and the Ministry of Social Justice & Empowerment to expand access to voice technology for people with speech, hearing, and voice-loss disabilities.

Commending the pace of AI adoption and positive disposition towards AI in the country, Chauhan pointed out that the only gap present in voice AI adoption was India’s multilingual and cultural width. “There are 22 national languages and many more dialects. We have some great partners here who are helping us obtain data sets for local languages, but traditionally the largest data sets for AI training were in English,” he stated. “We’re in the process of bringing all 22 languages and want to show a better, more efficient, and nuanced capture.”