Thank you, Isabel,
Mr. Shah, Mr. Mariano, dear all,
It is a pleasure to be part of this important discussion, bringing together technical experts and engineers working at the forefront of interoperability and AI.
I am particularly pleased to see the policy and technical communities coming together for this very timely conversation. Let me bring to you the development perspective of these issues.
With the extraordinary pace of developments in AI and the concerns being raised by the AI community itself, we have a collective responsibility to ensure that AI benefits are shared more equitably and that the risks are addressed effectively and without delay. This can only happen if we all work together to find the right answers and solutions.
I will speak to you today from a policy and development perspective and touch on two main themes.
First, data governance and AI governance, both of which are connected.
Ultimately, AI governance begins with data. How data is collected, accessed, shared and reused influences how AI systems are trained, how their outputs are assessed and verified, in which languages for which audiences, and ultimately it determines who benefits from the value they generate. We need to get the balance right.
Interoperability could serve as an enabling condition in ensuring that the benefits of data and AI governance are better shared. While we do not need to harmonize identical laws, technologies or regulatory approaches, we need to enable diverse systems to interact effectively while preserving legitimate differences in national priorities and regulatory frameworks.
We must work together to ensure that the information can be securely exchanged, consistently understood and effectively governed.
My second point is that we must consider the backdrop of persistent digital divides and increasing market concentration.
For UNCTAD, this is at the very core of our work. Developing countries risk remaining primarily providers of raw data and consumers of digital products, while capturing only a limited share of the value generated.
We already see significant disparities in the digital economy between developed and developing countries.
The digital economy is characterized by growing concentration of technological capabilities, infrastructure, data and market power in a small number of firms and countries.
By any of the measures we track in our research, technological capabilities, and AI in particular, are heavily concentrated in two global leaders. These two global leaders accounted for 40% of global corporate AI research and development, and for close to 80% of private investment in AI, 68% of top AI researchers and over two thirds of the top 100 companies in the field.
Furthermore, the gap is also visible in the ability of developing countries to capture value from the digital economy. In 2024, digitally deliverable services accounted for only 16% of services exports in least developed countries, compared with 61% in developed economies.
In this context, technology and interoperability are not neutral, and their benefits are not automatic.
Interoperability can reinforce existing asymmetries if the costs of adapting to standards fall disproportionately on countries and firms with fewer resources, or if the value generated from connected data and AI systems is captured primarily by actors with greater power.
Yet, if done right, greater interoperability can reduce fragmentation, facilitate innovation and lower barriers to participation.
For this reason, interoperability by design must also mean development by design.
This requires a coordinated response at the global level.
The United Nations has an essential role as an inclusive forum for developing shared principles and greater coherence across data and AI governance.
Global governance arrangements should support interoperability across jurisdictions while respecting national policy space and different levels of development and reducing the need and cost of navigating a fragmented landscape.
Global rulemaking and standard-setting must be genuinely inclusive. We need to consider affordability, competition and the avoidance of vendor lock-in from the outset.
Open and inclusive standards are important, such as those issued by IEEE and ISO, but they also need to be accompanied by investment, capacity-building, technology transfer and partnerships.
The goal is to enable developing countries and smaller firms not simply to connect to interoperable systems, but to participate meaningfully in their design and benefit from the value they create.
Internationally agreed principles must be translated into practical support, in the form of investment in digital infrastructure, institutional capacity and skills, as well as technology transfer and partnerships.
Let me conclude with one central message.
The success of interoperability should not be judged simply by how seamlessly systems connect, but by what those connections enable, for whom, and under what conditions.
That, ultimately, should be the measure of interoperability by design.
Thank you.