New Delhi: A set of common initiatives and directives signed by India, the US, China and 83 other countries at the recently concluded India AI Impact Summit will shape global collaborations on artificial intelligence (AI) despite the agreement being non-binding in nature, said a top bureaucrat who spearheaded the declaration.
On Saturday, the government released the New Delhi declaration that was signed by 88 entities, including the European Union (EU) and the International Fund for Agricultural Development (IFAD).
The declaration underlined the setting up of two platforms for global sharing of tools and principles in AI, as well as a global network of scientific institutes that pledged to collaborate in various areas of research using the technology.
In an interview with Mint, S. Krishnan, secretary in the ministry of electronics and information technology (Meity), said that the countries signing the declaration expressed their intent to adhere by the pledges, and ultimately it is adherence that matters.
“We should not be so concerned with the declaration being binding or non-binding. Ultimately, it is adherence to the declaration that will be important and relevant,” Krishnan said. “On this note, if nations voluntarily bind themselves to a declaration, this likely would make a bigger difference, than making a compulsion in some form.”
To be sure, this was the fourth global AI summit, with the previous editions held in the UK, South Korea and France, and the next year’s event slated for Switzerland. All the previous editions have produced non-binding policy directives in AI—with limited impact, thereby raising questions on their effective policy impact.
Prior to the summit, industry stakeholders had raised questions on whether the five-day event, the biggest of its kind on AI in India, would achieve a fruitful policy consensus. Kashyap Kompella of RPA2AI Research and Anushree Verma of Gartner, both veteran AI analysts, had told Mint that more than the policy consensus, the bigger takeaways will be around the focus on AI models, startup pitches, investments, networking opportunities and innovation.
“If you look at the previous editions of the summit, the declarations were not specifically directive in nature. There was the setting-up of AI safety institutes after the first Bletchley Park summit in the UK, but that did not lead to large-scale global collaborations or anything definite,” Gartner’s Verma had said. “This is a key problem you face when you don’t have a binding multinational agreement, something like the Paris climate accord.”
The secretary, however, maintained that a binding declaration would not necessarily signify success. “If you want a binding agreement between nations in international law, you have to sign treaties. And in today’s world, even with treaties, do you get the feeling that something is binding? The WTO (World Trade Organization) is a key example,” he said.
Policy stakeholders said the proof of the pudding is indeed in the eating—what happens next will be keenly eyed.
“On paper, the idea of a global network of state-funded scientific research institutions working together in frontier AI is great. Given that state-funded research projects are often more constrained than their private counterparts, a global sharing of resources could give them the ability to pursue wider work areas. The question, though, is if it would really play out in the way the New Delhi declaration has intended,” said Rohit Kumar, founding partner at The Quantum Hub, a technology policy consultancy firm.
Ashish Aggarwal, vice-president of public policy at India's technology industry body Nasscom, said that the New Delhi declaration was largely in line with India’s AI policy approach so far. “Prima facie, this can help create a policy directive and allow Indian startups to work for global markets as well. But in the end, it would all boil down to whether this actually plays out or not,” he said.
The AI Impact Summit concluded on Saturday, with information technology (IT) minister Ashwini Vaishnaw stating the event drew over 500,000 people to New Delhi’s Bharat Mandapam. Startups and top executives of global technology firms hailed it for being the first event of its scale that allowed small startups in AI to find clients, investors and partners.
The summit, however, did not come without its string of setbacks.
On 16 February, chaos descended upon the 125,000 early visitors—who, along with exhibitors, were stuck with lack of clear communication on organizational logistics. A theft of ₹25 lakh at a startup stall further marred the inaugural day, with foreign delegates raising complaints as well.
On the second and third days, Greater Noida-based Galgotias University plagiarizing an exhibit led to their removal—with the institution adamantly denying any voluntary wrongdoing. Many exhibitors also said that while five full days of expo exhibition were advertised, at least two-and-a-half days were wiped out due to security strictures, leading to disappointment.
To this, Krishnan said that despite the early hiccups, the first edition fared strongly.
“The AI Summit met most of the expectations we had and even exceeded some of our own expectations and projections in terms of the number of people in attendance, and the sheer energy and enthusiasm with which people wanted to participate,” he said.
The AI Summit drew over $210 billion in data centre investments from various parties, as well as private investment commitments, such as a $1 billion one for AI startups in India over three years by the India DeepTech Alliance. OpenAI, world’s largest private AI startup, also announced a partnership with the Tata group for AI adoption.