Podcast Summary: This Week in Global Development

Episode Title: Global Progress in the AI Era — Governing the AI moment: From global dialogues to real development impact
Release Date: February 24, 2026
Host: Katherine (Kate) Warren (with Devex colleagues Adva Saldinger & David Ainsworth)
Guest: Amandeep Singh Gill, UN Undersecretary General & Special Envoy for Digital and Emerging Technologies

Episode Overview

This episode delves into the global challenge of governing artificial intelligence (AI) during a period of rapid innovation and rising adoption, especially in low and middle-income countries. Host Kate Warren is joined by Ambassador Amandeep Singh Gill—one of the chief architects of international AI governance at the UN. The discussion covers the current landscape of fragmented AI dialogues, the role of diverse actors (including AI labs, the private sector, and governments), and the path from high-level discussions to tangible impact on development, equity, and inclusion.

Key Discussion Points & Insights

1. The Fragmentation and Connectivity of Global AI Governance

[01:16 – 04:26]

• Current State: Recognition of AI's transformative potential is nearly universal, but approaches to governance are highly fragmented.
• Fragmentation Exists In:
  • Regulatory methods
  • Technology development
• Positive Trends:
  • Emerging global mechanisms like the Global Digital Compact, International Scientific Panel on AI, and international dialogues.
  • Efforts at capacity building and addressing the "AI divide" are ongoing though mechanisms are still being shaped.
• Gill’s Perspective:

  "There is now almost universal recognition that this is an important technology ... But ... there is a lot of fragmentation, both in terms of governance/regulatory approaches, but also in terms of technology development." — Amandeep Singh Gill [01:42]

2. Is a Single AI Governance "Operating System" the Goal?

[03:05 – 04:03]

• Gill’s Position:
  • A single global framework is not feasible right now; instead, establish interconnected, norm-informed, multiple operating systems.
  • Human rights, solidarity, and the SDGs should guide these systems:

    "We cannot have a single operating system today. ... The human rights norms ... the sustainable development agenda ... those will be important regardless of ... the operating system we end up with." — Amandeep Singh Gill [03:18]

3. The UN’s Convening Role & Multi-Stakeholder Engagement

[04:03 – 07:40]

• UN’s Unique Position:
  • Most inclusive body for global dialogue, now advancing broader participation from civil society, tech, and the private sector.
• Multi-Stakeholder Engagement:
  • AI has driven deeper and more effective stakeholder participation (compared to earlier, more intergovernmental dialogues).
  • Participation of independent scientists, open source communities, and civil society is critical both for substance and implementation.

    "We need the perspectives of these communities ... we cannot implement what is discussed in international forums without their full participation." — Amandeep Singh Gill [05:58]

4. AI Labs & Tech Companies as New Development Actors

[07:40 – 11:04]

• AI Labs as De Facto Development Players:
  • Companies like OpenAI and Anthropic are forming partnerships with development agencies, influencing global challenges.
• Risks & Opportunities:
  • Tech's focus is on product-market fit and "solving problems," but true development is about empowering people—expanding agency and choice.
  • Need for robust, ongoing conversations between private tech and development actors to surfacing potential harms (exclusion, bias, energy/water use, risks to children/women).

    "Often, in the tech companies thinking ... they think of this as solutions for solving specific development problems. ... Technology empowers, but it also disempowers, it includes, it also excludes ..." — Amandeep Singh Gill [08:24]

5. Accountability Mechanisms for AI

[11:04 – 14:28]

• In-House Governance:
  • Company initiatives like Anthropic’s "long-term benefit trust" are early moves; accountability mostly happens reactively post-crisis.
• UN's Aim:
  • Engage labs to clarify emerging patterns of governance, connect to best practices, and diversify participation in innovation.
  • Need to avoid concentration of innovation in a few regions (“monoculture doesn’t scale”—diversity is key).

    "We need more diverse participation in innovation. ... In nature, forests scale, but monoculture doesn't scale." — Amandeep Singh Gill [11:40]

6. Digital Public Infrastructure & the India AI Impact Summit

[15:21 – 20:39]

• India’s Role:
  • India has pioneered digital public infrastructure (DPI), building public trust and delivering real benefits (e.g., in healthcare).
  • The India AI Impact Summit is seen as a bridge: between global mechanisms and global south aspirations, and a testbed for AI applications at scale.

    "It's a huge lab for experimenting at scale with AI for social good. ... It's also frightening because then you have to do it carefully." — Amandeep Singh Gill [16:22]

• Replicability & Adaptation:
  • Each region—such as Africa—will chart its own course and priorities, for example, by focusing more on intra-regional trade and entrepreneurial enablement than on financial inclusion.

7. Balancing Innovation Speed and Public Safety

[23:23 – 27:05]

• The Challenge:
  • The private sector often moves fast ("move fast and break things")—this can outpace capacity of governments to govern.
• Solutions:
  • Address information asymmetries by bolstering public/academic sector understanding of AI’s capabilities and limits.
  • Develop open-source public technology, despite its risks, to ensure public interest is represented.
  • Multi-stakeholder, bi-directional governance design, informed by both the public and private sectors.

    "It could be easy ... to just sit down with the private sector, 'Tell us how should we do governance' ... but ... less is better. ... We need to inform it, but still keep the public interest in mind." — Amandeep Singh Gill [24:05]

8. The International Scientific Panel on AI: Purpose and Impact

[27:05 – 30:40]

• Panel Details:
  • 40 experts from diverse regions and backgrounds; first systematically/independently selected global panel of its kind.
  • Will produce evidence-based reports aligned with policy dialogue.
• Intended Impact:
  • Move from ad hoc to systematic, science-led AI policy.
  • Use panel recommendations to accelerate opportunity—“governance can be an accelerator of opportunity, not just a speed breaker for risk.”

    "Governance is not only a break on risk, but ... can be an accelerator of opportunity." — Amandeep Singh Gill [28:03]

9. Navigating Geopolitical Tensions and U.S. Reluctance

[30:40 – 33:56]

• UN Approach:
  • Emphasize humility, patience, and long-term thinking; avoid overreach amid skepticism.
  • Leverage engagement with a broad range of U.S. actors but recognize the importance of tangible, agile, innovative outcomes to counter narratives of bureaucratic inertia.

    "It's important for the UN not to overplay its hand ... it's a long-term institutional challenge for all of us." — Amandeep Singh Gill [31:18]

10. Red Lines & Hard Limits: Learning from Weapons Governance

[33:56 – 38:21]

• Parallel with Military AI:
  • Gill draws from his background on lethal autonomous weapons, stressing lessons learned on the importance of human accountability and international law.
• Development “Red Lines”:
  • Technology must not undermine hard-won progress on social justice, equality, or humanity; need to decide what AI actions to allow or forbid (in both civilian and military spheres).

    "We have a problem if technology takes us back ... Technology could just sweep away [centuries of progress] in no time." — Amandeep Singh Gill [34:37]

11. The Key Message: Who Benefits from This Technology?

[38:21 – 40:38]

• Gill’s Closing Thought:
  • Channeling Gandhi: Decision-makers should ask whether their actions benefit the most marginalized as a way to ensure AI is a force for good.
  • Reminds listeners to consider what kind of world we're leaving for future generations—are we enslaving children to technology or empowering them?

    "Whenever I'm thinking about something to do, I think of the most wretched person ... and ask myself, does that help that person?" — Amandeep Singh Gill [38:51]

Notable Quotes

• “We need the perspectives of these communities ... we cannot implement what is discussed in international forums without their full participation.”
  — Amandeep Singh Gill [05:58]

• “Technology empowers, but it also disempowers, it includes, it also excludes.”
  — Amandeep Singh Gill [08:24]

• “In nature, forests scale, but monoculture doesn't scale.”
  — Amandeep Singh Gill [11:40]

• "Governance is not only a break on risk, but ... can be an accelerator of opportunity."
  — Amandeep Singh Gill [28:03]

• "Whenever I'm thinking about something to do, I think of the most wretched person ... and ask myself, does that help that person?"
  — Amandeep Singh Gill [38:51]

Important Timestamps

• 01:42: Fragmentation in global AI governance.
• 03:18: Future of AI "operating systems".
• 05:58: Multi-stakeholder engagement’s necessity at the UN.
• 08:24: AI labs as powerful new development actors.
• 11:40: Need for diverse innovation and risks of monocultures.
• 16:22: India’s approach to digital public infrastructure as a bridge for the Global South.
• 24:05: Balancing regulation for safety with innovation.
• 28:03: Vision for the international scientific panel’s impact.
• 31:18: Navigating U.S. skepticism and the pace of global governance.
• 34:37: Lessons from military AI and establishing red lines.
• 38:51: Final reflection on beneficiary-centered technology, referencing Gandhi.

Conclusion

Ambassador Amandeep Singh Gill’s conversation with Kate Warren offers a nuanced roadmap for how international community, AI developers, governments, and civil society can collectively steer AI to maximize public good while safeguarding against risks. The journey from scattered conversations to coordinated action is ongoing—with the UN seeking to serve not as a controlling center, but as a convener, facilitator, and amplifier for diverse, inclusive, and development-focused AI governance. At the episode’s heart is a call for empathy, humility, and placing the most vulnerable at the center of decision-making in the rapidly evolving AI age.