Dan Banik speaks with Pedro Conceição, Director of the UNDP’s Human Development Report Office, about how the idea of human development continues to shape global thinking on progress and equality. Together, they explore the latest Human Development Report (A Matter of Choice: People and Possibilities in the Age of AI) and discuss how artificial intelligence can either widen divides or empower people to live fuller, freer lives.
In this episode of In Pursuit of Development, Dan Banik speaks with Pedro Conceição, Director of the United Nations Development Programme’s Human Development Report Office, about the enduring power and renewed urgency of the human development idea.
The conversation begins with the origins of the Human Development Reports, tracing how Mahbub ul Haq and Amartya Sen transformed a moral vision into a measurable framework that challenged conventional notions of progress. Pedro reflects on how this approach—anchored in expanding people’s choices and capabilities—remains vital in today’s polarized and uncertain world, where attention is often captured by crises rather than long-term human flourishing. Dan and Pedro discuss the latest Human Development Report, A Matter of Choice: People and Possibilities in the Age of AI, which examines how artificial intelligence is reshaping economies, societies, and the very notion of human agency. They explore both the promise and the peril of AI — how it can enhance learning, health, and livelihoods, but also deepen inequalities if access, bias, and control are left unchecked. The episode also touches on widening global inequalities, energy poverty in Africa, and the foundational investments (in electricity, connectivity, and education) required to ensure that AI serves as a tool for empowerment rather than exclusion.
Resources:
Â
Dan Banik:
Pedro, it’s lovely to see you again. Welcome to the show.
Pedro Conceição:
Thank you, Dan. A pleasure. I’m a great admirer of your show, and it’s an honor to be here.
Dan:
The last time we met was in New Delhi in January 2024. One thing I remember—apart from talking to you—is a short, 20-minute ringside chat I had with Meghnad Desai, one of the early champions of the Human Development Report, alongside Mahbub ul Haq and Amartya Sen. It was fascinating to hear him speak about how the concept grew from a small idea among friends into something so influential.
So, Pedro, in today’s polarized world—with geopolitical tensions, wars, famines, and genocides—how relevant is the concept of human development? When you travel and present the Human Development Reports (we’ll come back to those later), do policymakers still say human development is a priority?
Pedro:
Let me start by mentioning a recent book by historian David Engerman, Apostles of Development. He traces the history of Amartya Sen, Mahbub ul Haq, and others, and how ideas around human development emerged from South Asia—not from a detached academic exercise, but from lived experience. On that base, Sen’s capabilities approach provided the conceptual scaffolding for the simple yet powerful idea of human development.
To answer your question directly: human development—enabling people to live to their full potential—is not only relevant; I’d argue it’s more relevant than ever.
Dan:
I agree. I’m a huge fan of human development; Amartya Sen is my mentor. There’s nothing more important. Yet attention spans are shorter, and priorities often diverge between the Global North and South. You rightly noted the concept’s Southern origins. Many debates today hinge on the fact that some countries want more human development, while others are accused of not providing sufficient resources—think shrinking aid or wavering support for the 2030 Agenda and the SDGs.
In that setting, do you still hear policymakers say, “We’ve progressed on the HDI—moved from low to medium human development,” and so on? Does that happen?
Pedro:
It does. The Human Development Index (HDI) doesn’t capture the whole idea—it’s a relatively narrow metric—but it helps us track progress and compare trends across countries. One strength of the human development approach is that it pairs a compelling concept with measurable indicators.
Two headline findings from the latest report:
Dan:
Exactly what Meghnad Desai told me—and something Sen emphasized years ago: operationalizing human development through the HDI was clever because policymakers pay attention to rankings. While rankings can’t capture every dimension, they draw attention. Another important point that foreshadowed the SDGs is universality.
Pedro:
Right. The HDI was calculated for all countries. Attempts to split countries into fixed groups run counter to human development’s universal aspiration: enabling people everywhere to realize their potential. The SDGs reflect this same universality—unlike the MDGs.
Dan:
You’ve been to Norway many times. Whenever we’re not number one on the HDI, journalists ask, “What’s happening—aren’t we the best country?” I believe Iceland is currently number one and Norway number two. But for me, the more interesting story lies at the bottom of the list—South Sudan, Chad, the Central African Republic—where upward movement year to year really matters. The low and medium categories are especially interesting. Would you agree?
Pedro:
All categories are interesting—we have four. What’s notable is how countries “graduate” from low to medium, medium to high, and high to very high human development, alongside the global average improving. Over different time scales, you see examples like the Republic of Korea, Bangladesh, Vietnam, Indonesia, and, recently, Türkiye moving up. That attention to progression was exactly the intention of the index: not to make countries “Norway overnight,” but to incentivize steady movement up the ladder. And yes, movement can also be downward.
Dan:
Let’s talk about the current report. Everyone is talking about AI—in academia and policymaking, it’s the buzzword. Often, the focus here is on the negatives for democracy, governance, and elections—serious concerns. What I don’t see enough of, and what your report highlights, is how AI can be mobilized for human and sustainable development—for climate action, especially. How can AI be a force multiplier in education, health, and social protection? I know you’re also concerned about humans losing control over AI’s development—something I’ve discussed with Daron Acemoglu in relation to Power and Progress. I even saw a video of President Trump hosting tech leaders at the White House where the conversation was all AI. So: how can we use AI positively to promote human development?
Pedro:
First, why AI? We’re doing a trilogy of reports stemming from the 2022 HDR, which argued we’re living through a “novel” context of uncertainty driven by: (1) political polarization (last year’s theme), (2) rapid technological change—especially AI (this year’s theme), and (3) planetary change—climate and biodiversity (the 2026 theme).
Our report is about human development, not the technology per se. The headline is: we shouldn’t let the future be determined by technology or its capabilities. There’s a narrative that machines will acquire human capabilities and replace people. Acemoglu warns against deploying “so-so AI” to replace humans in tasks they already do well—yielding displacement without productivity gains, plus environmental costs (energy, water).
If we treat AI as a companion that augments human capability, opportunities abound—in education, health, and economic life. But to realize gains, organizations must adapt processes; AI isn’t plug-and-play. As Arvind Narayanan argues, AI is a “normal technology”: like previous waves, it requires reorganization before productivity gains materialize. Crucially, AI can amplify human creativity—scientific and artistic—if we use it wisely.
Dan:
It has to be human-led and iterative—inputs refined step by step. On efficiency: smart homes are fun, but what matters for human development are concrete examples—malaria protection, remote medical advice, access to learning materials. On climate, some place outsized hope in AI; others are skeptical. The real danger is losing control as the tech moves quickly. Can we still shape the trajectory, or is it a runaway train?
Pedro:
We’re not worried about “losing control” of the technology in the way some fear, and the report explains why. On AI for good: every week I see new applications. One recent example from PNAS involves detecting anemia—debilitating in many parts of Africa—via a photo of a fingernail, bypassing blood draws and labs. New applications like this are emerging constantly.
Ultimately, it’s a matter of human choice. If something “runs away,” it’s because of choices we make, not because technology dictates the future.
Dan:
The report has many compelling figures. One showed that by 2024, a lot of ChatGPT demand came from middle- and low-income countries, not only high-income ones. But investment in AI is concentrated in the U.S., China, and parts of Europe and the U.K. There’s a risk the technology caters to their needs and perspectives. Human development is universal; it must serve people facing floods in Bangladesh or extreme poverty and malaria—everywhere. I’m reminded of everyday apps: I started using Lifesum to track calories—useful, but imperfect. AI can identify items, but humans add nuance.
Pedro:
Exactly. On “hallucinations”: this is why we should be cautious about delegating high-stakes decisions—life, liberty, justice—fully to AI. Use AI as input, not as an automatic decider.
Second, bias is a reality. We cite analysis comparing large language model responses to the World Values Survey: model outputs correlate strongly with views in very high-HDI countries, and less so as you move down the HDI spectrum. Many values and perspectives are underrepresented. Still, for domains less culturally contingent—science, math, many health interventions—models are improving. The key is ensuring lower-HDI countries aren’t merely users but also producers of AI.
Dan:
Context matters—even in health. In the Global South, frugal, low-cost solutions may be optimal given constraints. Policymakers might choose the third-best option because it’s feasible. So we shouldn’t preach universal “optimal” fixes; we need context-specific ones.
Pedro:
Agreed—and that’s why humans must remain in the loop. The goal is to empower, say, a village health worker in India to serve more people with better information, not to bypass them.
One genuinely “ugly” trend: the decline in young people’s mental well-being in many countries. Historically, well-being across age followed a U-shape; now, in many places, young people report sharp declines—especially young women and girls—alongside the diffusion of digital technologies powered by AI. The causes and remedies are debated, but AI-enabled platforms are part of the story.
Dan:
This is a big issue in Norway too. Leaders were even asked to reveal their screen-time during the recent campaign. And it isn’t only a high-income concern.
Pedro:
Right. Among those with access, the intensity of social media use is much higher in low-HDI countries than the global average. From a development perspective, we should try to get ahead of this problem before it becomes universal.
Dan:
Colleagues and students in South Africa, Malawi, Ethiopia, and Kenya constantly raise energy poverty: 600 million people without electricity, 900 million using coal and firewood. Years of underinvestment in infrastructure mean AI alone won’t solve it; we still need power and connectivity.
Pedro:
I heard the same frustration during our consultations. Think of three revolutions: electricity, the internet, and now AI. Many feel they missed the first two and don’t want to miss the third. Electricity and internet are prerequisites; once you have them, AI access follows without heavy additional capital. That reality strengthens the case for investing in electricity and broadband for communities still lacking them.
Dan:
Final question. A recent UN meta-study suggested many UN reports aren’t widely read, despite downloads. Your report seems to be an exception. How often is the HDR downloaded, and how is it used in policy debates? Are people reading beyond the executive summary?
Pedro:
We obsessively track impact. Combined downloads and online views of the HDR typically reach hundreds of thousands annually—500,000 to 600,000. Our data portal gets roughly 1.25 million visits and downloads a year. We also track media references, academic citations, and—crucially—citations in policy documents like budgets and national development plans. Usage is not only broad but growing, and HDRs have long shelf lives. For context, compared with the World Bank’s World Development Report, our download numbers are about three times higher (by our estimates). We’re proud of the reach, but we’re never complacent.
Dan:
Please keep up the good work, Pedro—and congratulations on another excellent report. Some of my students have it on their syllabus, so we’ll help boost those downloads. Great to have you on the show.
Pedro:
Likewise, Dan. A pleasure—anytime.
Â