Back to summit speeches

Recovering Public Value for the Next Generation

H.E. Ambassador Stan Oduma Smith

Ambassador and Permanent Representative of The Bahamas to the UN.

…in the evolution of the international financial landscape, regarding skills, speed, scaling allocation, volume, calibrating parity across markets. All of this amid the fragmentation and the world-changing technologies that were affecting economic productivity in states. And in this discussion, there was a wondering about how to generate strategic systemic productivity in the global supply chain. Now, what they laid out were the institutional coordinates for how we got to where we are today. And I want to use that to position how we start this discussion. They referred to development and public goods first being provided simply by the state being given loans.

And then they transitioned, the World Bank and other institutions, to giving loans and giving expert consultants to support the loans for development guidance. And what they pointed out is that we are now in a space where new global assets are being discovered and new innovations are emerging. And in this emergence, the question is: How do states provide market resilience? And I want to shift from their discussion of how states provide market resilience to how we address, with the very same elements, considering skill, speed, allocation, and scale, a landscape of development for the common good, for public services, rather than by crafting value through market dynamics.

Now, the state has a mandate for cultivating well-being, addressing left-behind places, catalyzing its built capacity, and ultimately functioning as an effective insurer of last resort of national resilience. The brief's own evidence makes the point. In the 2026 OECD survey, assurance was the overriding issue, not adoption, for 36 countries that deployed artificial intelligence in government. This is not the first time a general-purpose technology has arrived faster than institutions are able to govern it.

The industrial age did not distribute its gains on its own. Whatever equity the industrial age achieved for now-developed societies, it was built deliberately and built late, through technical standards, a public capacity to inspect, and organized civic voice. Public services, we all know, are, by bureaucratic culture, fragmented across sectors of the state. Almost all of our artificial intelligence discussions in the marketplace today concern private-sector services delivered outward to the public.

The larger opportunity that I want you to consider is the capacity of the state to see itself. What I mean by that is this: public project overruns; procurement records that do not reconcile with delivery records; information about a single program that may sit across four ministries, with no one holding the whole of it; or competing institutional visions of the same commitment. Very little of this is the result of dishonesty. Most of it is an information problem inside the hypothetical administration we are discussing.

It has more moving parts than it has capacity to observe, or parts opaque by culture, not incentivized to optimize its capacity to observe and make monetizable, or collateralize as a public benefit, the needs of the public. And these scenarios are precisely what artificial intelligence systems are good at transforming into equity. The brief already places this within safe scope, under helping staff to find and use internal knowledge. Governments, by common sense, you will already appreciate, hesitate for political rather than technical reasons. The phenomenon would be that an administration would be looking harder at its own performance, producing evidence against itself.

But there are two design choices that can change this. The first design choice is where the task and purpose sit. It belongs with the institutions whose constitutional purpose is already to look at state institutional and resource effectiveness or efficiency: the Auditor General, the Public Accounts Committee, the public investment function in finance. The second design function and resource consideration is where the recovered value goes; that has to be addressed by any one of the constitutional institutions we just referred to.

Our interest here is not to create a new efficiency unit inside the executive, but to consider how to transform our times and existing office of accountability interests. Efficiency has no constituency while its proceeds are invisible. Let us consider this: suppose it was settled in advance that value recovered, savings, avoided overruns, or assets brought back into use, is paid into a dedicated fund with a published mandate and a purpose named for the next generation. Finding a problem then produces something to announce, and the audit finding becomes a deposit.

This is the logic national investment funds have used for a half century, prior to 2008 in Norway and then in Timor-Leste, for instance, and agreed under governance standards at Santiago since 2008. The resource this contribution addresses is the value a state loses to its own blind spots. In every country participating in this discussion, there is a considerable sum that is lost to these blind spots. We would take a resource governments attempted to consume and place it behind a mandate, a board, and public reporting.

I would ask this program to treat such public value sitting unrealized as a question for your own year-long process. Now, I have no instructions to nominate a public institution of the Bahamas to the first cohort. But artificial intelligence offers the non-distracting capacity to offer simulations of how any such exercise would illustrate the performance of unrealized public value. Such value should be held behind a mandate for citizens, rather than captured by whomever controls the cable infrastructure through which the state's digital richness flows.

I would add that whatever the state gets back, whether savings, overruns being avoided, assets put back to work, or earnings from public data, these should be assigned to a single fund with a published mandate and held for the next generation. Among any of our differing positions, we should agree that what we want to be involved in is creating terms that align technology initiatives to generating standards that enable the four framing mandates of state obligation to their publics mentioned at the start of my thoughts here: well-being, places left behind, advancing the built capacity of the state, and the dynamism of the state as insurer of last resort.

Second, we want systems that can turn institutional discovery tasks into assets, as referred to here. And embedded in such a culture has to be administrative design that can foster agile institutional adoption that reflects public need. Otherwise, this is the challenge, or rather the problem: technologies will continue to outpace regulatory oversight and intensify social stresses and burdens that programs like this are inspired to arrest. Thank you.