Read this first — what this event was
Executive summary
Pakistan and Tajikistan used an HLPF 2026 side event to put one proposition on the UN record: AI will either bridge the development divide or widen it, and the difference turns on the choices made now — above all, whether young people in developing countries are treated as innovators or left behind. The room was unusually unified. A Tajik minister, a Pakistani youth-program chairman, a UNDP advisor, a Google lead, a youth delegate, and floor interventions from Kazakhstan, China and Azerbaijan all converged on the same commitments.
Four commitments the whole room shared — and named as the closing takeaways:
- Bridge the AI divide — compute, finance, skills and data must reach developing countries. →
- Turn AI opportunities into tangible development outcomes, equitably. →
- Youth as shapers, not users — innovators and co-authors of AI governance. →
- Human-centred, safe, trustworthy AI — "humanity must shape the future of AI." →
Underneath the consensus run three tensions of emphasis — finance vs. skills, sovereign capacity vs. Big-Tech partnership, and state optimism vs. UNDP's risk warning — differences in weight, not open disputes.
Governing frame — where this plugs in
The reference point four speakers kept returning to
UN Global Dialogue on AI Governance + Independent International Scientific Panel on AIHeld in Geneva the week before this event, the inaugural Global Dialogue and its Scientific Panel were cited by UNDP, Google, China and Pakistan's Permanent Representative alike as the frame for turning AI principles into practice. The Panel's preliminary warning — that safeguards are not keeping pace and the access gap risks widening inequality — was quoted approvingly across the room. Pakistan's PR framed the global approach as currently "fragmented," with this dialogue as the integrator.
- No hard deadline emerged from this event; the actionable takeaway is which venue the agenda now runs through.
- Parallel venue China announced the 2026 World AI Conference and high-level meeting on global AI governance, Shanghai, 17–20 July 2026 ("Intelligent Partners: Co-Creating the Future") — offered days before it opened, a competing/complementary track to the UN Geneva dialogue.
Where they agree — the shared frame
- AI must bridge, not widen, the divide. The central risk everyone named: concentrated compute, finance, data and talent will deepen inequality without deliberate action.
AI is not inherently inclusive. Its benefits will depend on the choices we make today… artificial intelligence may widen the very inequalities that sustainable development seeks to overcome.
— Chairman Rana Mashood Ahmad Khan (Pakistan) - Youth are shapers, not just users. The organizing thesis of the whole event, echoed by China, Azerbaijan and the closing remarks.
History will not remember the speed of AI that we adapted towards, but history will remember how we had the courage to include the young people while we were adapting AI.
— Bisma Qamar (Pakistan, youth delegate) - AI must remain human-centred, safe and trustworthy. Transparency, human oversight, accountability.
We say that artificial intelligence will shape our common future, but… it's the humanity that must shape the future of the AI.
— Ambassador Asim Iftikhar Ahmed (Pakistan PR, closing) - AI can accelerate the SDGs — especially SDG 9. Innovation, resilient infrastructure, inclusive industrialization; named by Tajikistan, Pakistan, Azerbaijan and China.
For developing countries, the promise of AI lies not in automation alone, but in its power to leapfrog traditional development barriers.
— Deputy PR of China - International cooperation and capacity-building are the mechanism. Regional hubs, shared infrastructure, South-South and triangular cooperation — Tajikistan's IT Hub Dushanbe, Kazakhstan's Astana Hub, China's governance initiative, the UN dialogue.
Developing countries should not merely adopt artificial intelligence, they should actively shape its future.
— Permanent Representative of Kazakhstan
Who spoke — the voices in the room
Minister for Economy & Trade of Tajikistan
Co-host keynote · Government of Tajikistan
Tajikistan is first in Central Asia with a national AI strategy (to 2040); declared 2025–2030 the "years of the digital economy"; sponsored the July 2025 UNGA resolution on AI for sustainable development in Central Asia; is building the regional AI Center / IT Hub Dushanbe (construction from 9 June 2026).
"Ensuring equitable access to AI technologies, strengthening capacity building, and promoting international cooperation must remain at the center of our collective efforts."
⚠ Name unresolved on the recording (headed "Abdurrahman Zadeh," introduced as "Ramonov"); withheld pending verification.
Rana Mashood Ahmad Khan
Chairman · Prime Minister's Youth Program of Pakistan
Youth-first keynote (~70% of Pakistan is under 30). Catalogues the PM Youth Program — ₹60 bn for 2026–27, 1.6 M laptops, 73k AI-trained, National AI Policy targeting 2 M AI professionals by 2030 — and sets the four priorities. "Investment, not declarations"; AI must not become a "technological monopoly."
"The true measure of artificial intelligence will not be the sophistication of its algorithms, but the lives it improves."
Jennifer Lui
AI Trust & Safety Advisor, Digital AI & Innovation Hub · UNDP
The panel's counter-current. Compute concentration (¾ in one country); safeguards not keeping pace; governance made through distributed "micro-decisions." UNDP evidence: ~25 AI landscape assessments; RapidAssist post-disaster AI; the AI Hub for Sustainable Development (with Italy, 18 African countries). Warns of youth cybercriminality and weak, invisible procurement.
"How do we make sure the frontier benefits actually reach us?"
Hashim Saeed
Go-to-market Lead · Google
AI as "a bridge, not automation," on three pillars — startups, enterprise, academia — toward "localized, sovereign solutions," delivered via Google programs (AI Seekho: 1 M professionals by 2030; Build with AI in 100+ countries; the AI Sprinters Framework for localized datasets; Gemini/NotebookLM piloted in Sindh & Punjab ministries). Synthetic-ID watermarking and agent "observability" as governance tools.
"The true power of AI should not be restricted into the hands of the few, but should be intelligence for the many."
Bisma Qamar
Youth delegate of Pakistan · PM Youth Program
Youth as contributors to policy, not observers; the barrier is platforms and funding, not talent. Cites the National Youth Council (100 members), Digital Youth Hub, Young Innovation Awards, Green Youth Movement. Calls for AI-ethics education — the "fine line between exposure and exploitation."
"Let's choose inclusion over exclusion when it comes to ideas and innovation."
Kazakhstan
Permanent Representative · Government of Kazakhstan
National showcase: 10th globally on the Online Services Index, 24th on e-government development; two NVIDIA AI supercomputers; a piloted national LLM; Astana Hub; founding member of global AI capacity-building networks; the new Almaty Asia-Pacific Digital Solutions Center. "AI should unite rather than divide."
"Developing countries should not merely adopt artificial intelligence, they should actively shape its future."
China
Deputy Permanent Representative · Government of China
Leapfrog framing: intelligent manufacturing +30% efficiency / defects halved (SDG 9); digital public services to 90%+ of rural populations; 10,000+ AI startups; national AI curricula; data-protection & algorithmic-accountability law; contributes to the Scientific Panel. Announces the Shanghai World AI Conference, 17–20 July 2026.
"AI can reduce rather than widen the digital divide, provided that the infrastructure investment and the capacity building reach the margins."
Azerbaijan
Representative · Government of Azerbaijan
Praises the co-hosts' youth-driven model. National AI strategy 2025–2028; Digital Development Council (58 initiatives); an AI academy and research lab; the "My Government" e-gov platform and a national data center; a comprehensive legal reform defining AI, cybersecurity and technology transfer in law for the first time.
"So that no country and no young person is left behind in this AI-driven economy."
The Aung Hwan
Delegate from Myanmar · public-policy student, U. Wisconsin–Milwaukee
The conflict-state edge case. Since the 2021 military coup, Myanmar's digital development "deteriorated." Asks how vulnerable, marginalized youth in conflict-affected and least-developed countries can access, adopt and innovate with AI — pressing "leave no one behind" to its hardest case. Went substantively unanswered.
"How can we address the vulnerable and marginalized young people, especially from these conflict-affected countries or least developed nations…?"
Qasim
Neural Lab · AI consulting firm
The procurement-access question: how can young entrepreneurs actually work with UNDP and the multilateral space as contributors, not just recipients of frameworks? Jennifer Lui's candid answer — even UNDP hasn't codified its AI procurement, and that opacity is a real barrier.
"Not just as a recipient of this framework, but also as a contributor that can help us develop a sustainable solution."
First name / firm only on the recording.
Advait
Founder, "Govinform" · USA
Shifts the youth conversation from access to motivation: young people are excited about AI's possibilities "but a lot of times they don't know how or have little motivation" to engage. How to build a clear path in?
"How can we make it so that the youth that want things to change but don't know how… how can we incorporate them in the whole process?"
First name only on the recording; organization name unverified.
Subhan Malik
Audience
Q directed to Tajikistan: building on the theme of inclusive development, how can we ensure the AI revolution does not further widen the divide between developed and developing nations?
"How can we ensure that the AI revolution does not further widen the digital divide between developed and developing nations?"
Ambassador Asim Iftikhar Ahmed
Permanent Representative of Pakistan to the UN · closing remarks
Distils the discussion into four priorities (divide, outcomes, youth-as-shapers, human-centred AI); calls AI a global public good; notes the global approach is "fragmented" and looks to the Geneva Global Dialogue + Scientific Panel to integrate it.
"It's the humanity that must shape the future of the AI."
Saima
Moderator · Prime Minister's Youth Program of Pakistan
Frames the event, sequences the speakers, and sets the recurring "youth as shapers, not users" refrain between statements.
"Young people must not merely adapt to the revolution; they must be empowered to shape and lead it."
First name only on the recording.
Tensions of emphasis — differences in weight, not open disputes
1 · Finance vs. skills — what actually bridges the divide?
Is the binding constraint money and compute that must move, or skills and capacity that countries can build?
The two emphases sit side by side without confrontation. The Pakistan chairman makes the sharper structural demand — investment and compute must move, and declarations are not enough. Most other speakers stress the more consensual, unilaterally actionable lever: skills, scholarships, academies, curricula.
"The global AI divide cannot be bridged through declarations alone. It requires sustained investment, equitable access, and genuine international cooperation."
"Only 14% of workers have actually globally received formal and true AI training. So we have to bridge this gap through academia and enterprise."
"Funding is almost the biggest challenge that youth face" — the barrier is platforms and access, not talent.
2 · Sovereign capacity vs. Big-Tech partnership
How do developing countries build sovereign AI while scaling through a single dominant vendor's tools?
Structural rather than voiced. Google frames the path as a transition to "localized, sovereign solutions" — while positioning its own products as the delivery mechanism. The Pakistan chairman names the guardrail directly. Neither treats the other as an opponent, but the sovereignty-vs-monopoly question is live in a room where a Big-Tech partner is presenting.
"A transition from passive consumption of global technologies to active, localized, sovereign solutions." (delivered via AI Seekho, Build with AI, Gemini in ministries)
"AI must remain human-centered and locally relevant. It must not become a tool of technological monopoly."
3 · State optimism vs. UNDP's risk warning
Is the story mostly national AI wins to celebrate, or risks that are outrunning the safeguards?
The floor interventions largely catalogue national successes. Against that optimistic register, UNDP's Jennifer Lui is the one voice pulling the other way — the compute gap, safeguards falling behind, and the darkest note of the day on jobless youth and cybercrime.
China: "Our intelligent manufacturing systems have boosted efficiency by over 30% while cutting defects by half… digital public service platforms now reach more than 90% of rural populations."
"Our safeguards are not keeping pace with AI's growth. And the gap in access — in compute, in skills, in infrastructure — risks exacerbating the inequalities into the next generation."
Stand-out interventions — the ones worth reading twice
The compute-concentration + safeguards warning — Jennifer Lui, UNDP
The event's most substantive departure from the celebratory tone. Concrete UNDP evidence: ~25 AI landscape assessments worldwide; RapidAssist post-disaster AI delivering 72-hour assessments in 20+ countries (350,000 people reached in Afghanistan after the 2025 earthquake; redeployed in Venezuela); the AI Hub for Sustainable Development (with Italy) supporting 18 African countries on green compute. Her governance thesis is the intellectual high point: AI governance is "being decided in many smaller choices… micro-governance choices you're making, whether or not you call it that."
The conflict-state edge case — The Aung Hwan, Myanmar
The one voice from a conflict-affected / least-developed context. Since the 2021 coup, Myanmar's digital development "deteriorated." Presses "leave no one behind" to its hardest case — vulnerable youth in conflict zones — and, tellingly, drew no substantive answer in the panel's closing round.
The procurement-access admission — Qasim (Neural Lab) → Jennifer Lui
Asked how startups can actually contract with UNDP and the multilateral space, Lui was unusually candid: even UNDP has not codified its AI procurement practices, they are "not visible," and that opacity is itself a barrier — "I will also admit that even our own organization here is still kind of working through what those procurement practices need to be."
China's parallel-venue pitch — Deputy PR of China
Beyond the national showcase, China used the floor to announce the 2026 World AI Conference and high-level meeting on global AI governance (Shanghai, 17–20 July 2026, "Intelligent Partners: Co-Creating the Future") — a parallel governance track offered days before it opened, alongside the UN's Geneva dialogue.
Talking points — the quotes to carry
"AI is not inherently inclusive. Its benefits will depend on the choices we make today."Chairman Rana Mashood Ahmad Khan (Pakistan)
"The true measure of artificial intelligence will not be the sophistication of its algorithms, but the lives it improves. Not the wealth it generates for a few, but the opportunities it creates for many."Chairman Khan (Pakistan)
"Of the world's most powerful computing infrastructure, three-quarters of it sits in only one country… the central question is not how do we build the next frontier of AI. It's how do we make sure the frontier benefits actually reach us?"Jennifer Lui (UNDP)
"Our safeguards are not keeping pace with AI's growth."Jennifer Lui (UNDP), citing the Independent Scientific Panel
"More youth now, especially in Global South countries, are being enlisted into cybercriminality. Why? Because they don't have other jobs and because they are digitally literate."Jennifer Lui (UNDP)
"We have to treat AI not as a tool for automation, but as a bridge for opportunity… a transition from passive consumption of global technologies to active, localized, sovereign solutions."Hashim Saeed (Google)
"History will not remember the speed of AI that we adapted towards, but history will remember how we had the courage to include the young people while we were adapting AI."Bisma Qamar (Pakistan, youth delegate)
"Developing countries should not merely adopt artificial intelligence, they should actively shape its future."Permanent Representative of Kazakhstan
"AI must not become a tool of technological monopoly."Chairman Khan (Pakistan) — the sovereignty guardrail
"We say that artificial intelligence will shape our common future, but it's the humanity that must shape the future of the AI."Ambassador Asim Iftikhar Ahmed (Pakistan PR), closing
- Bridge the AI divide — "that should remain our central task."
- Turn AI opportunities into tangible development outcomes — equitably.
- Recognize young people as innovators, decision-makers and partners in shaping AI governance — not simply users.
- Keep AI human-centred, ethical, transparent, safe and accountable — quoting UNDP: "the safeguards are not keeping pace with the growth of AI."