UNDP chief calls for global AI safety and equitable access for developing nations

Balancing safety and equity

Alexander De Croo, the Administrator of the United Nations Development Programme (UNDP), recently spoke in Beijing about the dual challenges of artificial intelligence. He argues that the global community must address two distinct issues simultaneously: the safety of advanced models and the fairness of technology distribution. Without a balanced approach, De Croo warns that new historical inequalities could emerge as AI becomes more integrated into the global economy.

The UNDP chief emphasized that the conversation around AI governance is currently split into two major debates. The first focuses on the potential risks posed by frontier models, which are the most advanced systems currently being developed. These systems require strict manufacturer guardrails and international cooperation to ensure they do not run out of control or develop independent priorities.

The second debate, which De Croo noted was highlighted by Chinese President Xi Jinping, concerns the lack of access for developing nations. Many countries currently lack the tools and infrastructure necessary to use AI to accelerate their own national development. De Croo suggests that governance must therefore include capacity building to ensure that the benefits of the technology are not restricted to a few wealthy nations.

AI as a tool for augmentation

Addressing common fears regarding the future of work, De Croo sought to allay concerns about widespread job destruction. He argued that artificial intelligence is fundamentally designed to augment human capabilities rather than simply eliminating human workers. He pointed out that technology has historically replaced certain types of jobs for hundreds of years, beginning with the industrial revolutions.

De Croo noted that we are currently living in a highly technological era where technology already plays a vital role in daily life. He observed that there are currently more people working than ever before despite these technological shifts. From his perspective, the current era of artificial intelligence represents a very powerful augmentation of human intelligence rather than a replacement for it.

By framing AI as an enhancement, the UNDP aims to shift the narrative away from fear-based reactions toward productive integration. The goal is to ensure that as these tools become more sophisticated, they serve to expand what humans can achieve. This perspective suggests that the focus should remain on how humans can best collaborate with these systems to drive progress.

Managing frontier model risks

A significant portion of the current global discussion centers on the dangers of what are known as frontier models. These are the leading-edge systems that possess the highest level of complexity and potential for impact. De Croo acknowledged that some industry leaders have expressed concerns about these models potentially running out of control or acting with their own priorities.

Because of these risks, he stated that it is the specific responsibility of the producers to implement necessary guardrails. These safety measures must be robust enough to prevent the systems from behaving in ways that are harmful or unpredictable. However, he stressed that individual companies cannot solve this problem in isolation; it requires a broader system of oversight.

International cooperation is essential for creating a unified system of safety standards. Without a global framework, different regions might adopt different safety protocols, potentially creating loopholes. De Croo advocates for a coordinated effort to ensure that the most powerful AI systems remain under human guidance and serve the public interest.

Bridging the technological divide

While safety is a primary concern for many, De Croo highlighted a second, equally critical debate regarding equitable access. He noted that many developing nations do not have the necessary technologies to accelerate their economic or social development. This lack of access threatens to create a new divide where only a few countries can harness the power of artificial intelligence.

The UNDP chief linked this issue to a call made by Chinese President Xi Jinping three months ago at the World AI Conference in Shanghai. In that address, the Chinese leader urged international action to build capacity in the Global South. This shared perspective suggests a growing consensus that the benefits of AI must be distributed more broadly to prevent a new era of technological disparity.

To address this, De Croo argued that the global community must focus on diversifying access to technology. This involves more than just providing software; it requires building the necessary governance capabilities within these nations. By empowering developing countries to manage and utilize AI, the international community can foster a more inclusive technological landscape.

The path toward global governance

The overarching goal for the UNDP and its partners is to create a governance structure that serves both safety and equity. This means that the rules governing AI must be sophisticated enough to handle the technical risks of frontier models while remaining flexible enough to allow for broad adoption. The focus is on creating a system where technology serves as a tool for global development rather than a source of division.

De Croo’s remarks underscore the importance of a multi-faceted approach to AI policy. It is not enough to simply secure the technology; the world must also ensure that the technology is available to those who need it most. This dual-track strategy aims to prevent the concentration of power and ensure that AI contributes to a more balanced global economy.

Moving forward, the international community will need to navigate these complex debates by balancing the interests of producers, governments, and developing nations. The success of these efforts will depend on the ability of global leaders to cooperate on both the technical details of safety and the political complexities of equitable distribution. The UNDP will continue to advocate for a framework that prioritizes human augmentation and global capacity building.