A Slowdown Imperative: The UN’s AI Warning in Context

The United Nations’ call for global cooperation on artificial intelligence has taken on an increasingly urgent tone. Secretary-General António Guterres warns that “the world cannot afford a race to the bottom on AI safety,” echoing concerns raised by experts and industry leaders about unbridled AI growth.

Guterres emphasizes both the enormous potential benefits of AI and its substantial dangers, making it clear that collective action is necessary to mitigate risks. However, obstacles to a coordinated effort are multifaceted and complex. Domestic and global competition, particularly from China, hinders progress, as do profit motivations driving companies like Meta to distance themselves from calls for a slowdown.

Even industry leaders who have sounded the alarm about AI’s risks – such as Elon Musk, Sam Altman, and Dario Amodei – face pushback from those who view efforts to limit AI as part of some “conspiracy.” The recent warning from the General Assembly’s 40-member panel on AI highlights real-world risks, including harm to mental health, potential use as a destructive tool, and impacts on social, economic, and environmental systems.

Guterres’ call for global coordination is not about stifling innovation or limiting AI benefits. Rather, it aims to establish guardrails that prioritize human dignity, transparency, accountability, and safety. This requires more than national action; international cooperation is essential, with countries willing to exchange information, agree on common standards, and establish trust among nations and stakeholders.

The UN’s proposal for a slowdown in AI development has precedents, such as the Paris Agreement and the Montreal Protocol. However, these initiatives often stalled due to competing interests and conflicting priorities. Guterres’ warning should be seen as a clarion call for leaders to put aside their differences and work towards a shared understanding of how to advance AI safely.

The next week’s high-level session of the U.N. General Assembly will be a crucial test of leaders’ commitment to addressing AI risks. While Guterres has expressed reservations about holding an impromptu summit, he remains convinced that a shared understanding of safe AI development is essential. The stakes are too high for anything less than a collective effort as the world hurtles towards increasingly powerful systems.