A major step in voice technology has emerged as Josh Talks introduced what it describes as the world’s first full-duplex conversational AI model in Hindi. The system is designed to listen and speak simultaneously, enabling voice interactions that closely mirror natural human conversation.
Unlike traditional voice assistants that work in a turn-based format—where users must finish speaking before receiving a response—the new model can process speech while replying at the same time. This allows the system to acknowledge users mid-sentence, respond instantly, and handle interruptions without disrupting the flow of conversation, the company said.
“Voice AI has learned to recognise speech and generate speech. What it has not yet learned well is how to participate in conversation. This research shows that when models are trained on large-scale natural dialogue, they can begin to learn the rhythm of how people actually speak to each other,” said Shobhit Banga, Co-founder of Josh Talks.
“By building this infrastructure using Hindi conversations, we have demonstrated that full-duplex conversational systems can be developed beyond English,” he added.
The AI model, called Human-1, was trained using 26,000 hours of natural two-person Hindi conversations from 14,695 unique speakers. According to the company, the dataset captured real conversational elements such as overlapping speech, interruptions, pauses, acknowledgements, and spontaneous reactions. These factors are crucial because real-life conversations rarely follow strict or scripted patterns.
To support further research, Josh Talks has open-sourced the Hindi duplex conversational model. However, the underlying dataset used to train the system will remain proprietary. The release allows researchers and developers to experiment with full-duplex dialogue systems and build new applications based on the approach.
While the current model focuses on Hindi, the company said the method can be adapted for other languages. With India’s vast linguistic diversity, large-scale conversational datasets could eventually help enable natural voice interactions across multiple Indian languages.
By focusing on authentic conversational behaviour, Human-1 aims to move beyond the limitations of traditional voice systems and better adapt to the way people naturally communicate, the media edtech company said.
Also read: Viksit Workforce for a Viksit Bharat
Do Follow: The Mainstream LinkedIn | The Mainstream Facebook | The Mainstream Youtube | The Mainstream Twitter
About us:
The Mainstream is a premier platform delivering the latest updates and informed perspectives across the technology business and cyber landscape. Built on research-driven, thought leadership and original intellectual property, The Mainstream also curates summits & conferences that convene decision makers to explore how technology reshapes industries and leadership. With a growing presence in India and globally across the Middle East, Africa, ASEAN, the USA, the UK and Australia, The Mainstream carries a vision to bring the latest happenings and insights to 8.2 billion people and to place technology at the centre of conversation for leaders navigating the future.
