UN chief stresses global cooperation on AI safety, says concerns cannot be ignored amid calls to slow development

United Nations Secretary-General Antonio Guterres has called for stronger international cooperation to ensure artificial intelligence is developed safely and responsibly, warning that the rapid expansion of AI is moving faster than the global understanding of its potential risks. In a post on X, Guterres reiterated concerns he had raised earlier this week, saying the world cannot ignore warnings coming from people involved in building increasingly advanced AI systems. He stressed the need for a coordinated global effort to make AI more transparent, accountable and safe, while keeping human dignity at the centre of its development.

“AI has enormous potential,” Guterres said, while noting that concerns are growing among those working directly on the technology about the pace at which it is advancing. His remarks come as the technology industry continues to debate whether the development of more powerful AI systems should be slowed or accompanied by stronger safeguards.

Speaking at a press conference at the opening of the 81st session of the UN General Assembly on Wednesday, Guterres said artificial intelligence could deliver major benefits in areas such as sustainable development, education, healthcare and climate resilience. At the same time, he warned that the potential risks require serious attention. Guterres said some experts have called for a pause in AI development, while others have pushed for stronger guardrails. He argued that concerns raised by people working at the forefront of AI research should not be overlooked.

The UN chief also placed responsibility on national governments, saying they must protect citizens from emerging risks linked to artificial intelligence. He stressed that domestic action alone would not be enough because AI is a technology whose impact extends across national borders. According to Guterres, international coordination will be essential to deal with these challenges. He cautioned that voluntary measures implemented differently across countries may fail to provide a consistent response to the risks posed by rapidly developing AI systems.

The debate has intensified following Anthropic CEO Dario Amodei’s public call for a slower pace of AI development. The issue has also attracted attention from other leading technology executives, including OpenAI CEO Sam Altman, Microsoft CEO Satya Nadella and Google’s Demis Hassabis. Concerns have also emerged among AI researchers themselves. Jacob Coxon, who was associated with Anthropic, announced his departure from the industry while raising concerns about companies moving quickly to create increasingly capable AI systems that could become difficult to control.

Against this backdrop, Guterres said AI governance should be guided by scientific evidence. He highlighted the UN’s Independent International Scientific Panel on AI, which brings together experts from around the world to help ensure that policymaking is based on research and evidence. Guterres said artificial intelligence would be among the issues discussed with world leaders and indicated that the United Nations would continue working on practical approaches to managing emerging risks.

He also called for a common international framework to determine when increasingly powerful AI systems should face additional safeguards or other risk-management measures. The secretary-general linked the AI debate to broader global challenges, including the climate crisis and growing inequality. He argued that international cooperation is becoming even more important as countries deal with multiple interconnected threats.

The discussion over AI development is therefore moving beyond the technology sector and increasingly becoming a matter of international policy. Guterres’ latest remarks underline the growing push for countries to work together on safety standards while continuing to explore the potential benefits of artificial intelligence.

Disclaimer: This image is taken from X/@antonioguterres.