UN Chief Urges Global AI Cooperation, Warns Against Safety 'Race to the Bottom'

United Nations Secretary-General António Guterres has called on global competitors in artificial intelligence to work together on the threats posed by the technology, cautioning that "the world cannot afford a race to the bottom on AI safety."

Speaking to reporters at the UN, Guterres said concerns voiced by those working on the frontlines of AI — that its development "is racing ahead of our understanding of the risks" — cannot be dismissed.

He argued that governments bear a primary duty to shield civilians from AI-related dangers, making "national action essential," while stressing that "global coordination is also indispensable." Guardrails, he said, are needed to build trust and to ensure AI remains safe, transparent and accountable, with human dignity at its centre.

The remarks came in response to a proposal by Anthropic CEO Dario Amodei to coordinate a slowdown in AI development among leading technology companies and governments. In recent days, OpenAI's Sam Altman, SpaceXAI's Elon Musk and other prominent industry figures have said the pace of AI development needs to slow. Meta CEO Mark Zuckerberg, however, distanced his company from that position, saying every AI firm must ensure its technology is safe.

Other hurdles to a coordinated approach include domestic and global competition, particularly from China, and profit motives. There has also been pushback from President Donald Trump, who this week described threats that AI might destroy humanity as "a HOAX" and characterised efforts to limit the technology as a "SICK conspiracy."

Guterres has acknowledged AI's enormous potential — to accelerate economic development and education and to strengthen health systems — but has also warned of the "existential" threat of "runaway artificial intelligence." He said AI would be a major topic in his discussions with world leaders at next week's high-level session of the UN General Assembly.

"The United Nations will continue to explore practical ideas and approaches in the immediate days ahead," he said, adding that a shared understanding is needed on how to advance safe, secure and responsible AI development while identifying when increasingly powerful systems may require stronger safeguards or additional risk-management measures.

At Guterres' request, the General Assembly established the first global scientific body on AI in 2025. In its preliminary report in July, the 40-member panel warned that rapid, unchecked use of AI at scale risks harming users' mental health, "potential use as a destructive tool, impacts on social, economic and environmental systems, and challenges associated with controlling the technology."

The secretary-general said he is aware that geopolitical divides could complicate efforts, but argued that countries "that have the highest capacity in the frontiers of AI" should establish contact, exchange information and agree on common guardrails "to avoid this race to the bottom that could lead one day to a gigantic disaster at global level."

Asked whether next week's gathering of top officials from the UN's 193 member nations could serve as a summit on curbing AI excesses, Guterres said there was not enough time to prepare such a meeting properly. He added, however, that "it makes full sense at the right moment to have a serious international summit on artificial intelligence," and that at such a summit "it is absolutely essential to have everybody on board" — governments, companies, scientists and civil society.