India's Generative AI Moment: Poised for Global Impact

Earlier this month I had an opportunity to visit India. While meeting close family, I was asked to mediate a family legal dispute involving unpaid debts. As I am not a lawyer by profession and have very limited knowledge of the Indian legal system, I thought of asking my friend "chatGPT" for advice. Indeed, I got an answer, but the terms were too generic to help resolve the situation. Remaining optimistic, I looked for an alternative and stumbled upon LawBotPro - a generative AI bot designed and built in India specifically for Indian laws. It helps provide legal guidance and answers questions. To my surprise, it answered my queries satisfactorily, offered mediation options citing relevant Indian laws (e.g. the Indian Penal Code), and enabled me to successfully resolve the family dispute.

This article is an inspiration from this “incident” on how India is considering Generative AI and what the future lies for India for Generative AI.

About India:

For those who doesn’t know India well, it has world’s 16% of population with highly diverse 1.37 billion people (where 47% of them are under the age of 25). While Hindi is the national language and English is widely spoken in the business world, Indian constitution has recognized 22 major language out of over 1,000 language spoken (It’s like Europe where EU recognizes 24 language amongst the 27 countries). India is the world’s fifth-largest economy by nominal GDP and third-largest by purchasing power parity (PPP). India accounts for 7.2% of the global economy in 2022 in PPP terms, and around 3.4% in nominal terms in 2022

While nearly 70% of India's GDP is driven by domestic consumption, the Information Technology (IT) and Business Process Outsourcing (BPO), or IT-BPM sector, which contributes heavily to India’s economy, makes up only 7% of the GDP.

About Generative AI:

With recent advancements in AI/ML and adoption of Generative AI, IT-BPM sector will be impacted the most and so does their productivity as well as revenue. Those who are still relatively new to Generative AI, here is a primer:

  • Generative AI is a technological shift where system is generating content, whether it’s text, image, audio, video, or even code for the software development.

  • By applying generative AI, an individual or a corporation can gain massive amount of productivity, to begin with.

  • For example, if it takes 8 hour for a developer (for the IT segment) to write certain lines of code, with the help of generative AI tools, they may be able to complete it within couple of hours or less.

  • Similarly, the customer service agents who answers the phone/manage chat (for the BPO segment) and interact with customers would be able to answer quickly (either by searching the information quickly, AI produces the answers before even it reaches the agent, etc.) - improving the productivity anywhere between 30 to 60%.

Coming back to India:

Generative AI has huge potential to transform industries across India. While exports stand to gain tremendously in the IT-BPM sector, the domestic application of this technology will likely have an even greater impact.

Just as LawBotPro localizes legal chatbots for India, we can expect to see customized generative AI solutions emerge for every major sector - from agriculture to real estate, retail to financial services, insurance to gaming and sports. As developers build bots leveraging data and knowledge specific to the Indian context, businesses small and large can harness the power of AI to drive efficiency, personalization and innovation. Whether it's providing personalized crop advice to farmers, making insurance accessible to the underserved, or enabling startups to scale customer engagement, India-centric generative AI can uplift livelihoods and living standards nationwide. The possibilities are endless, but localized applications are key to unlocking AI's full potential.

The potential of localized, multi-lingual generative AI in India is exemplified by innovative services like KissanAI for the agriculture sector. KissanAI blends GPT-3.5 and Whisper models to empower farmers with customized advisory in 10 regional languages. This is game-changing for Indian agriculture. Imagine an aspiring young farmer lacking experience to make informed choices - KissanAI can guide them on optimal seeding rates, fertilizer application and more, truly democratizing domain knowledge.

For centuries, people in India, including myself, have believed that most solutions to life's problems can be found in religious scriptures. Thus, reading these texts daily or deriving answers from them is thought to help improve one's daily life. To assist with this, a religious foundation has created "GitaGPT" based on the Bhagavad Gita, considered the most auspicious scripture in Hinduism. GitaGPT allows users to interact with it like a chatbot to receive insights and guidance in English. Though currently limited to English questions, were GitaGPT made available in 20 other regional Indian languages, use of such services could spread tremendously.

Regional language support:

While chatGPT and other AI operate primarily in English, initiatives like Nandan Nilekani's AI4Bharat (AI for India) are building open foundation models to advance generative AI in regional Indian languages. AI4Bharat has developed state-of-the-art models across 22 languages for tasks like language translation. For example, their Anuvaad model is translating Indian legal documents at scale. As an open source project, models like Anuvaad could soon empower services like LawBotPro to allow users to address legal queries in their regional local language. This exemplifies the efforts underway to make generative AI more accessible by incorporating Indian languages, contexts and use cases.

Looking forward:

Enhancing literacy, increasing awareness of rights, and expanding technology access are crucial to overcome barriers to advancement in India. It is hard to overcome because of people are comfortable using their regional local language and far more comfortable writing and speaking in that language. For these reasons, large number of government schemes, rights and entitlements do not reach the last mile. With the advent of foundation models we are on the cusp of the transformation to empower over a billion citizens in wide variety of domains - justice, health, education.

While services like KissanAI, GitAI, and LawBotPro already exist in pockets, widespread adoption of homegrown, generative AI that is accessible to all and targeted to local needs can drive impactful transformation.

McKinsey’s estimates of global GDP increase anywhere between $2.2 Trillion to $4.4 Trillion will likely exceed even the most optimistic projections, with India at the forefront of this transformation. Combining this with the youth population who are ready to adopt the technology - remember 47% of population is 25 or younger - India is going to make even bigger impact to the domestic and global economy.

Previous
Previous

Last week in AI - week40-2023

Next
Next

The Cost of AI: Should You Build or Buy Your Foundation Model?