Amandeep Singh Gill, recognized in this year’s Artificial Intelligence Strength Index, is part of a small group of crucial leaders responsible for shaping how artificial intelligence is integrated into global systems and international diplomacy itself. As UN Under-Secretary-General and Special Envoy for Digital and Emerging Technologies, Gill faces one of the most complex policy challenges of our time: how to ensure that artificial intelligence enhances humanity rather than destroys it. Gill is candid about the assumptions, power dynamics, and governance barriers that define the global AI landscape. He challenges the common notion that AI will automatically provide “infinite abundance” and warns that AI-driven concentration of power and wealth in the hands of a few could erode human agency and freedom.
But he is not a pessimist. For Gill, recent declines in training costs for smaller language models and signs that scaling laws may be reaching their limits bode well for a potentially more democratic, less centralized future for AI. From negotiating frameworks for harmonizing disparate national AI capabilities to addressing the ethical and humanitarian dangers of autonomous weapons, Gill believes governance guardrails are an important foundation for innovation. “Innovations that undermine trust and security can be toxic to anyone who cares about artificial intelligence,” he said. Gill’s work reflects the UN’s broader ambition to promote responsible development of AI, build digital capabilities, foster cooperation and invest in shared human and technological infrastructure.
One assumption about artificial intelligence that you think is completely wrong?
Artificial Intelligence will usher in an era of infinite abundance by transforming humanity Turn intelligence into a commodity. “Intelligence is too cheap to count” reminds me “Electricity is too cheap to meter” was a phrase from the heyday of nuclear hype.
If you had to pick one moment in the last year when you thought, “Oh, crap, this “Changing Everything About Artificial Intelligence”, what is it?
The fall in training costs occurred almost simultaneously Smaller and more efficient LL.M. (DeepSeek, Indus, WinAI) and Ilya Sutskever NeurIPS talks about scaling laws reaching a steady state. it gives me hope Different long-term outcomes are possible, namely more democracy, less A centralized, more sustainable AI innovation ecosystem.
What keeps you awake at night most in AI development? What are people not talking about?
The Social and Political Consequences of High Centralization of AIPower and wealth are in the hands of a few. I worry about the growing gap between China and the United States Architects of cognition and those who inhabit cognitive spaces It undermines human agency and freedom.
What are the biggest challenges facing countries with huge differences in AI? The ability and philosophy to agree on a common framework and how you Balancing innovation and security?
This is a collective action problem familiar to us in other areas such as climate Change. At the United Nations, we address this problem through a multidimensional, differentiated approach The agenda not only addresses the different priorities of countries; Stakeholders including the private sector – promoting AI use, risks, governance interoperability (of interest to the private sector), capacity building (especially concerns of developing countries), etc. We also rely on agreed norms, e.g. Harmonizing perspectives on human rights. The innovation versus security dilemma is wrong. Innovation always works best within guardrails. Innovations that undermine trust and security poison everyone who cares about AI
How the United Nations is using artificial intelligence to accelerate progress towards sustainable development Development goals, especially those of developing countries? What exactly is AI? What apps have you seen that have the biggest impact on technology bridging? gap?
Our focus is on improving the nation’s ability to responsibly develop and deploy artificial intelligence for the Sustainable Development Goals. The ability to develop solutions rather than solution. Artificial intelligence will have beneficial impacts in countless fields, Deliver government services, planning and monitoring in real time Infrastructure projects, agricultural extension and livestock/fisheries governance, green transition, public health and diagnostic equity, Education etc. We’ve only scratched the surface of some applications, Examples include data-driven AI tools for predicting food insecurity. secretary–General’s latest report on innovative financing solutions for artificial intelligence capacity building The case is made for urgent investment globally in key enablers, e.g. as computational, cross-domain talent and context-rich datasets. extensive Innovation bases can unleash AI solutions close to demand.
How to deal with the tension between military artificial intelligence applications and artificial intelligence applications? What are the humanitarian concerns in your diplomatic work and how is that going? An international agreement on lethal autonomous weapons?
From a UN perspective, we remain concerned about the potential of military AI Application that undermines respect for international humanitarian law, sparks new arms race and lowering the threshold for conflict. The Secretary-General believes Life and death decisions cannot be entrusted to machines and calls for UN member states should establish clear bans and regulations on such systems by: 2026.
If you were tasked with using AI to protect students in schools, what approach would you take and what key challenges would you need to address? How can AI contribute to prevention and early intervention efforts without creating a surveillance environment that disrupts the educational experience?
I would let students think about the pros and cons of AI as a tool to address their insecurities. For example, if they believe bullying is a problem or there is a risk of gun or knife violence, I ask them to think about how AI tools can help identify victims or perpetrators and prevent intimidation or violence. This will also enable them to consider the trade-offs between data collection and use or artificial intelligence versus simulation solutions.
When thinking about this question, I was reminded of an experiment by researchers at Carnegie Mellon University in which air quality sensors were introduced into a classroom. Students experimented with them, taking them to parking lots and found that pollution peaked during pick-up and drop-off times. They used the data they collected to engage in conversations with parents and change behaviors to improve air quality. This is an example of empowering the use of technology, which is how we should approach AI solutions in general.
