MeitY: India’s AI and Deep Tech Ecosystem Should Focus on Patience, Not Quick Profits

by Mimansa SidhnathTalking about patient capital, Additional Secretary to the Ministry of Electronics and Information Technology (MeitY), Abhishek Singh said that one of the major problems in the Indian AI industry today is the short supply of patience required in the process of AI capability building, rather than plain capital. “The problem is not of capital per se, the problem is of patient capital. Most investors—most HNIs [High Net-Worth Individuals], most VCs [Venture Capitalists]—in India look for return in a very short gap.

Whereas in deep tech and AI, the investment cycles are longer. One has to wait for years in order to get returns,” he said. He illustrated his point with the example of NVIDIA, a major industry player at the moment, saying, “Even for a company like NVIDIA, they had to invest for 30 years before it started seeing results and that is seen across the ecosystems.



” Singh was speaking at the 3rd Indian consultation round of UNESCO’s AI Readiness Assessment Methodology (RAM) in Hyderabad on April 8. RAM is one of the two key tools (the other being Ethical Impact Assessment) outlined by UNESCO in its Recommendation on the Ethics of Artificial Intelligence, an “actionable” framework for the ethical development and use of AI, which was signed by 193 member states, including India, in 2021. The first consultation was held in New Delhi and the second in Bengaluru, Karnataka.

Singh mentioned that the next consultation is set to take place in Guwahati, Assam. Ikigai Law is acting as the implementation partner of the initiative. GDP Share in AI Building and Identifying Key PlayersSingh further pointed out that the other factor that contributed to complication of capital availability was the GDP share going into AI development.

“Second problem is that Research and Development (R&D) expenditure has been much lesser compared to the global averages. For example, our R&D budget is 0.7% of our GDP whereas in China and the US, it is about 3% of their GDP,” he said.

He went on to list some recent initiatives by the Government to “address this problem” like the Anusandhan National Research Foundation (ANRF). The Foundation has a corpus of Rs 1 lakh crore for the next 5 years to enable the taking up of R&D projects, he explained. Additionally, the Government also announced an outlay of Rs 10,000 crore for deep tech startups, he reminded.

Furthermore, under the IndiaAI Mission the Government is allocating approximately Rs 2,000 crore for direct startup funding, with another Rs 2,000–3,000 crore to be deployed indirectly for supporting compute infrastructure and foundation model development.“If we equate the total..

.this is anywhere between Rs 35,000 to Rs 40,000 crore available from the Government this year. And given that this is through Fund of Fund (FoF) model, it can actually unlock a much greater amount.

So funding is available today. What we need to do is how quickly we can deploy, how quickly we can create a framework, working with investors, VCs, incubators, and accelerators to be able to identify which are the startups that we can bet on. We see a lot of traction, even though there’s a narrative about India being behind in deep tech startup,” he said.

Techno-Legal Approach: The New Phrase in the TownOn being asked about the governance mechanisms the Centre is building to ensure AI systems are ethical by design, Singh reiterated his past stance that the Government was looking to regulate AI applications and not the tech itself. In this vein, he pointed towards the recently published draft AI governance guidelines. “That document proposed a techno-legal approach to AI regulation that defines how regulation will be carried out.

It was published for public consultation. We have got more than 120 odd comments and we are finalising the document based on the feedback we have received,” he remarked.The techno-legal approach seems to be the latest buzzword used by officials.

Earlier in the month, Union Minister for Electronics and Information Technology, Ashwini Vaishnaw also alluded to this approach when talking about AI regulation, explaining that this approach helps law enforcement to harness the potential of tech experts like researchers and scientists.”Law alone will not be sufficient,” Vaishnaw had said, ringing familiar to Singh’s statements. Meanwhile, the AI Governance Guidelines in question are driven by a principle-based approach and keep harm mitigation as its core principle.

This becomes complicated given that the Government is planning to refrain from introducing any AI-specific laws. “Primarily, we are not looking at drafting a new law for regulating AI. We do believe that the existing legal framework, the IT Act, the Intermediary Guidelines, and the Bhartiya Nyaya Sanhita, are sufficient provisions for ensuring that harm that can result from AI can be regulated,” Singh said.

As pointed out by the Internet Freedom Foundation (IFF), “these principles [on which the Guidelines are based] often remain purely aspirational if not backed by clear enforcement mechanisms.”States Will Implement AI Mandates, Not the CentreSingh pointed out that the States have emerged as major stakeholders in drafting guidelines around AI since “the Government of India [the Central Government] has a limited role when it comes to building AI applications and technology solutions. Ultimately, the adoption will happen in the States.

” He explained this is where RAM comes in, and that these consultative workshops are helping the Centre connect with the stakeholders, like startups, researchers, Government officials, and get their perspective with regard to how the guidelines are to be framed. “And once we are able to capture that, then only whatever we are prescribing as an assessment methodology or as a regulatory framework or guidelines will be successful.” Singh also said that to build “safe, trusted, responsible, ethical AI” under RAM, “the Government is supporting projects in research institutions and startups” to build tools that ensure that AI models and applications that are being developed conform to the principles of RAM.

“[By] just creating a framework, or guidelines, or passing a law, we will not be able to do much,” he said.Since the session with the Additional Secretary was not open to questions, here are a few unanswered questions: How major a role would BNS (erstwhile Indian Penal Code), as mentioned by Mr. Singh as one of the laws that would govern AI in India, play in AI regulation? How is the Government planning to limit the misuse of the law? How is the Central Government planning to support States in the implementation of RAM and other AI mandates? Will the Government make public the comments it has received on AI Governance Guidelines? Mimansa is a freelance journalist and researcher.

Also Read:Govt is Looking to Regulate AI Applications, Not AI Technology: Additional Secretary to IT MinistryAnalysing MEITY’s Report on Development of AI Governance GuidelinesHow Can AI Regulation Mitigate Harm Without Stifling Innovation? #NamaThe post MeitY: India’s AI and Deep Tech Ecosystem Should Focus on Patience, Not Quick Profits appeared first on MEDIANAMA..