AI companies race to adapt chatbots to India’s many languages

Global tech companies as well as local start-ups are looking to tap deeper into the Indian market with artificial intelligence (AI) platforms adapted for India’s vast range of languages, Financial Times wrote.

India has 22 official languages, with Hindi the most widespread, but researchers estimate the languages and dialects spoken by its 1.4 billion people rise into the thousands, according to the article.

Also Read | Satya Nadella stepped in to solve big AI ‘chip feud’ between Nvidia and Microsoft over this issue

Examples of companies and their products

Microsoft, Google, and start-ups including Silicon Valley-backed Sarvam AI, founded just last year, and Krutrim, founded by Bhavish Aggarwal of Indian mobility group Ola, are all working on AI voice assistants and chatbots in languages such as Hindi and Tamil, the article read.

The tools are aimed at fast-growing Indian industries, such as the country’s large customer service and call centre sector, according to the article.

Google launched its Gemini AI assistant in nine Indian languages on Tuesday.

Microsoft’s Copilot AI assistant is available in 12 Indian languages, and the company is working on other projects tailored for India, including building “tiny” language models at its Bengaluru-based research centre.

These can run on smartphones rather than the cloud, making them cheaper and potentially better suited to countries like India where connectivity can be limited, the article read.

Microsoft is also partnering with Bengaluru-based Sarvam AI, which is developing generative AI tools for Indian businesses. The start-up raised $41 million from investors including Peak XV, Sequoia’s former India arm, and Menlo Park-based Lightspeed Venture Partners, according to the article.

Also Read | The AI-Fintech connection: Transforming the financial ecosystem

The background

Investing in local AI companies is becoming more important as governments seek to develop “sovereign AI” that is trained and stored within their borders, Hemant Mohapatra, partner at Lightspeed India told Financial Times.

The details

India’s AI race does not involve building LLMs (large language models) from scratch to compete with leaders such as Open AI, with investors arguing the resources and capital required would be too much, the article read.

Instead, companies such as Sarvam AI are focusing on adapting existing LLMs for Indian languages and using voice data instead of text, making them more effective in a country where many prefer to communicate through audio messages rather than in writing, according to the article.

Also Read | Nvidia's Jensen Huang praises Ilya Sutskever who left OpenAI: ‘Began AI revolution’

Other benefits

An additional benefit of testing new technologies and tools in a country of India’s size and diversity is that they could be exported elsewhere, Tanuja Ganu, a manager at Microsoft Research in Bengaluru, told Financial Times.

“It’s using India as a test bed and validating some of the technology in India and seeing how we can expand it to other parts of the world,” she said.