Global Leaders Unite in Delhi to Forge AI Safety Cooperation Framework

"The race to build ever more capable and powerful AI systems is accelerating, with AGI and advanced robotics on the horizon, and safety is clearly not the priority. This needs to change," said Nicolas Miailhe, Co-Founder of AISC, in his opening remarks.

"For the very first time, we are building technology that could become more intelligent than us and that we don't understand," he added.

The day's programme explored five core themes through nine panel sessions, special addresses, workshops, and demonstrations of practical AI safety tools.

India's approach to AI safety remained a central focus throughout the convening. Former India G20 Sherpa and ex-CEO of NITI Aayog Amitabh Kant delivered a special address, framing AI as a transformative force that must remain equitable.

"If AI is not utilised by vast segments of populations, if it is not utilised by women, by farmers, and if it is not utilised by students across the world, then AI is not fit for purpose," said Kant.

He pointed to India's experience with digital public infrastructure as a model, noting that "battles for new technology are never won by the first mover" and arguing that India's linguistic diversity and scale position it to shape how AI serves the next five billion people moving from poverty to the middle class.

The most pressing global safety risks were examined across multiple sessions, with speakers stressing the gap between the pace of AI development and the international response. Dr Eileen Donahoe, Founder of Sympatico Ventures and former US Special Envoy for Digital Freedom, challenged the framing of safety as an obstacle to progress. "In some circles, both in DC and Silicon Valley, AI safety and security has sometimes been characterised as an impediment to innovation. I think that framing is backwards," said Donahoe.

"Broad societal adoption and global diffusion won't happen if governments, enterprises, consumers, and citizens don't trust in the basic reliability and safety," she added.

Donahoe emphasised that the most urgent AI safety and security risks are inherently transnational, and the solutions must be transnational.

The role of middle powers in shaping global AI governance emerged as one of the day's most significant discussions.

Netherlands Prime Minister Dick Schoof delivered a special address arguing that countries beyond the AI superpowers are far from powerless.

"We are not powerless spectators watching from the sidelines. We, the middle powers, represent the largest part of the world economy and the strongest democratic traditions. Together we form the majority," said Prime Minister Schoof.

He outlined the interdependence of the global AI supply chain and called on middle powers to seize the moment: "It's up to us to take a responsible approach, to write a different story. One of cooperation, not confrontation; of shared standards, not fragmentation; of democratic values, not profit maximisation."

International coordination mechanisms were scrutinised across multiple sessions, with participants examining how to translate principles into operational governance. Lucilla Sioli, Director of the European AI Office, outlined the EU's emerging approach to facilitating compliance through voluntary codes of practice, noting that 27 companies have already signed up. "AI, as we know, has no borders," said Sioli.

"I really hope that the discussions at this summit will look at cooperation in AI governance and how the experience of different countries can be used for a possible framework of the future," she added.

Sioli highlighted that the European AI Office would soon publish research in the journal Science on proportionate, risk-targeted AI evaluation methods.

The urgency of moving from principles to action was underscored by Dr Andrew Forrest, Founder of Minderoo Foundation, who delivered a forceful call for regulation in his opening address.

"Releasing the power of AI without regulating its clear and obvious danger is absolutely irresponsible on a historical scale," said Forrest.

"You can't manage what you can't measure. Our moment in history is to ensure that the immeasurable becomes managed, that the measurable becomes regulated." He called on leaders across business, politics, and the social sector to come together to ensure that the potential harms of AI are checked with the same seriousness as its benefits are pursued.

The day also featured a fireside chat between Turing Award laureate Professor Yoshua Bengio and Skype founding engineer Jaan Tallinn on the trajectory of frontier AI, with Bengio warning that "in a matter of years, not decades," AI systems will be capable of most cognitive tasks humans can perform. Panels with representatives from Microsoft, Google DeepMind, Amazon Web Services, and the Frontier Model Forum examined industry approaches to safety, while parallel sessions showcased live demonstrations of AI safety evaluation platforms, child safety testing systems, and risk modelling frameworks.

"The word safety keeps missing from the title -- and, we fear, from the agenda -- of these conversations," said Cyrus Hodes, Co-Founder of AISC. "The goal of AI Safety Connect is to put it back."

AISC's engagement at the India AI Impact Summit continues with Shared Responsibility: Industry and the Future of AI Safety on 19 February, examining private-sector approaches to safety coordination, and International AI Safety Coordination: What Policymakers Need to Know on 20 February, a ministerial-level panel during the main Summit proceedings.

- ANI