UN Bets on Science to Steady AI Governance
At the India AI Impact Summit, the UN urges fact-based AI governance, warning that without scientific consensus regulation risks fragmentation and delay.
Topics
News
- India Plugs Into Pax Silica as AI Supply Chains Realign
- India AI Impact Summit 2026: Did India’s Tech Push Just Take a Hit?
- When Autonomous AI Tests the Limits of Oversight
- UN Bets on Science to Steady AI Governance
- Legacy IT Leaders Say India’s AI Future Lies in Talent, IP
- PM Calls for Human-Centric AI at India Summit
Global leaders, industry chiefs and policymakers converged at the India AI Impact Summit 2026 in New Delhi this week to debate how science can guide AI governance, with UN Secretary-General António Guterres calling for fact-based policy ahead of a planned global governance dialogue.
“If we want AI to serve humanity, policy cannot be built on guesswork. It cannot be built on hype or disinformation,” Guterres said, warning that technology is advancing faster than governments can interpret it. Politics without evidence, he argued, risks swinging between paralysis and overreaction.
The United Nations has established an Independent International Scientific Panel on Artificial Intelligence, a 40-member body meant to provide a shared analytical baseline before a Global Dialogue on AI Governance in July.
AI systems cross borders effortlessly but regulation does not. Without coordination, Guterres cautioned, countries risk fragmenting into incompatible standards. The panel should ensure policy is “neither a blunt instrument that stifles progress, nor a bystander to harm.”
Yet the challenge lies as much in science as in geopolitics.
Yoshua Bengio of Mila Institute noted that researchers “don’t always agree” on future risks or even how to interpret current evidence. Unlike climate modelling, AI lacks decades of stabilized data. Capabilities advance unevenly, outperforming humans in narrow tasks while failing in others.
“There is going to be a lag,” he warned, “because things are moving too fast for policy to catch up.”
That lag complicates risk assessment. Policymakers may have only “clues” about severe outcomes. Act too early and they face criticism; act too late and they invite harm.
Brad Smith, President of Microsoft Corp., turned to institutions. Eighty years after its founding, he argued, the United Nations remains indispensable.
“There is no such thing as a crystal ball,” he said, dismissing sweeping forecasts of imminent upheaval. The real question is “how will we use those machines to make people smarter?”
Soumya Swaminathan, former chief scientist at the World Health Organization, compared AI governance to early covid pandemic policymaking, when evidence shifted rapidly and guidance evolved.
“Policy must change. But ask for the relevant evidence,” she said, stressing adaptability over speed.
Geography also matters. Balaraman Ravindran of IIT Madras noted that most research on AI’s social effects comes from the West.
“We don’t have enough evidence about how AI is going to affect the social fabric” in India, he said.
Without local benchmarks, countries such as India risk importing assumptions that may not fit their social fabric.
Anne Bouverot of the French government framed the debate as one of understanding.
Quoting Nobel laureate Marie Curie, she said, “Nothing in life is to be feared. Everything is to be understood.” Whether AI destroys jobs or transforms them will shape whether governments prioritize income support or large-scale reskilling.
India’s principal scientific adviser, Ajay Kumar Sood, outlined a “techno-legal” approach, embedding governance into system design itself, drawing on lessons from digital public infrastructure.
The summit exposed a central paradox. Science is necessary for AI governance, but it cannot settle political trade-offs over risk, equity and economic priorities.
Guterres closed with a plea: “Less hype, less fear. More facts and evidence.” Whether that discipline can keep pace with the technology it seeks to shape remains uncertain.
