The boom created by OpenAI’s chatbot, ChatGPT, is still a topic of discussion as it has become a part of various industries worldwide. Meanwhile, the Indian billionaire and reliance industry owner Mukesh Ambani has taken a historic step that has led India into the competitive AI race. We have already witnessed that Ambani’s Seetha Mahalaxmi Healthcare has recently partnered with the reputed IIT Bombay and eight other Indian Institutes of Technology to launch “Hanooman”, a collection of large language models which is rigorously trained in 22 Indian languages. This BharatGPT initiative aims to democratise AI by bridging the linguistic gap and making it accessible to Indian users.

Mukesh Ambani unveiled Hanooman on 1st February 2024. The Indian AI model is marked as a potential rival to ChatGPT. This announcement marked a significant step towards India’s AI aspirations, which aim to address the concerns and limitations of existing language models and cater to the unique linguistic landscape of our country. It can respond in 11 Indian languages, including Hindi, Tamil and Marathi, and they are planning to expand it to more than 20 languages as well. Hanooman is predominantly designed to work in four fields: healthcare, education, governance and financial services.

The multimodal capabilities

Hanooman adds more to the table than text, unlike the traditional language models. The multimodality of Hanooman aids it in generating and analysing content in several media, which includes text, speech, and video. Hanooman assists in seeking healthcare advice in the native or local language, receives personalised video responses and produces educational content tailored to different learning types.

Open-source model

The first wave of Hanooman models with parameters from 1.5 billion to a starting 40 billion will be unveiled soon. It is significant to highlight that these AI models will be open-source and allow developers, academics, etc, to build on them and create specific applications for different areas.

The linguistic diversity

India presents a pivotal challenge for AI development with its diverse languages. This RIL-supported Hanooman addresses the concerns by assisting in 11 Indian languages such as Hindi, Tamil, Telugu, Marathi and Malayalam. This model’s goal is to integrate all 22 Indian languages, revolutionising access to information and communication for many people who struggle with communication in English while also helping them to use AI in their language.

This focus of Hanooman into the Indian languages is expected to bridge the digital divide and make AI more accessible to a wider population in India. This model’s capability is expected to accelerate innovation across several industries, which leads to new products, services and solutions. The model can also contribute significantly to building Indian language processing technologies, which is expected to benefit research and education.

Applications

The impact of Hanooman goes beyond the usual translation services. The chatbot offers seamless customer help service in local languages and empowers healthcare, governance, education and financial services. With Hanooman, India will be on the verge of a ground-breaking AI revolution that offers inclusivity, innovation, and empowerment.

The introduction of Hanooman ensures the first step for an AI model that is India-centric. However, the journey has just begun. The success of the model will be analysed by tackling the challenges, cooperating with stakeholders and constantly adapting to meet the varying needs of users and industries. It’s important to acknowledge that the AI model is in progress; hence, the capabilities will be more precise in the future. 

• Photo by Rohan Solankurkar on Unsplash