|
Listen to this story
|
In her Budget speech, Finance Minister Nirmala Sitharaman announced the establishment of three Centres of Excellence (COEs) in Artificial Intelligence intending to inch closer to the ambitious goals of ‘Make AI in India’ and ‘Make AI work for India’.
Along with the declaration of ‘Make AI in India’, the government has recently pushed new regulations for data usage. India also suggested boosting organisations like the National Data Analytics Platform and the Data Governance Quality Index during a recent meeting of the G20 Presidency working group. In the same speech, Sitharaman also announced the National Data Governance Framework Policy, which was released last year.
Vishal Gupta, founder and CEO, Seclore, believes that the focus on AI COE’s in combination with the National Data Policy and data embassies announced by the government will help companies address their security concerns related to AI adoption. These initiatives by the central government are indicative of the government’s interest in moving towards new technologies. But, what comes between intent and actualisation is the question: Does India have the potential, patience and infrastructure for AI to be developed at the home turf?
What the numbers say
In a written reply to a question in the RajyaSabha, Minister of State for Electronics and Information Technology, Rajeev Chandrasekhar citing the NASSCOM data, said that AI employment in India is estimated at about 416,000 professionals. However, as per the government data, sectors like IT services and BPO/ITeS, which play a huge role in the Indian economy, have direct employment of around 5.1 million (FY 2021-22).
With the indirect job creation of 12 million, AI professionals form a mere 2.4% in the entire tech ecosystem in India. While the government believes that the growth rate for the sector is around 20-25% and AI is expected to contribute an additional $957 billion to the Indian economy by 2035, the whole ‘Make AI in India’ movement is far from becoming a reality in the region.
According to Ratan Dargan, co-founder & CTO at ThoughtSol Infotech, India lacks the necessary resources to invest in the development of AI technologies such as skilled data scientists, infrastructure, and access to high-quality datasets. Additionally, Dargan believes that the lack of a unified regulatory framework for AI and its applications also poses a challenge. “The lack of public awareness and understanding of AI technologies and its implications may hinder its adoption,” says Dargan.
Further, he says, “To ensure success, we must continue to invest in developing our AI ecosystem, datasets, funding, and programming languages, as well as improving collaboration between industry, academia, and government to build a strong foundation for our AI capabilities.”
And this might be closer to reality. For instance, OpenAI took almost seven years before it was able to launch the near successful program ‘ChatGPT’. And not just time, it took billions of dollars in funding to be able to carry on the research. So far, OpenAI has received $12 billion in funding from various institutions with Microsoft leading it with the funding of $11 billion. This amount is equal to the GDP of Goa and twice to that of Tripura.
But, not all is lost
However, Chirag Gupta, managing partner of 8X Ventures, believes that “the cost of training AI models has been dropping over 50% per year, and data quality has improved drastically.” He believes that with its massive internet population, India is now sitting on a unique opportunity with the right data, training models and AI talent to build solutions better than ChatGPT in less than five years.
Additionally, to get on the same level of development in the sector, the Department of Science & Technology is implementing the National Mission on Interdisciplinary Cyber-Physical Systems (NM-ICPS).
The Indian Institute of Technology (IIT) Kharagpur has established Technology Innovation Hubs (TIH) on Artificial Intelligence and Machine Learning as part of this Mission with the goal of delivering cutting-edge training and building capacity for the development of next-generation scientists, engineers, technicians, and technocrats in the field of Artificial Intelligence.
And while government support is welcome in the field of AI development, it should not be limited to just relying on platforms already built by the world– something which India is known for doing. For instance, BharOS, which was built on the Android platform. The space program of India, although successful, goes on to apply the same principles. India skipped the entire computer revolution and leaped to be the country one of the largest smartphone users. India skipped the BS5 engines and from BS4, went straight to BS6 engines used internationally.
Are investors ready?
However, can we really do the same with Artificial Intelligence? Because to go from GPT 2 to GPT 3.5, OpenAI had to collect a tonne of data, investing a huge amount of money. The AI model has to be trained on previous generations before it can move forward. So, is the government ready to invest the required amount and time? Or will it be another ‘Make in India’ scheme where smartphones are assembled rather than being manufactured? And are Indian investors ready to wait for longer durations than they are usually familiar with for product development?
According to Gupta, “Launching exponential innovation in the AI industry often requires patient capital. India has made massive strides in nurturing the DeepTech VC ecosystem and is poised to disrupt the AI industry in the coming decade.”
