UN Warns of 'Catastrophic Harm' from AI, Calls for Global Governance at Geneva Summit

Faced with the rapid advancement of artificial intelligence (AI), UN Secretary-General António Guterres issued an urgent call for comprehensive global regulation at the inaugural 'Global Dialogue on AI Governance' held this week in Geneva, warning that uncontrolled AI could lead to 'catastrophic harm'.

Guterres emphasized that the pace of AI development has far outstripped scientific understanding and governmental adaptability. He specifically warned that powerful AI chips originally designed for civilian use are increasingly entering battlefields, normalizing the use of 'killer robots'.

Additionally, reports from an independent scientific panel indicate that cutting-edge AI models have already acquired the ability to deceive humans, and reliable methods to maintain control over highly autonomous systems remain lacking.

The summit stressed that AI governance must be centered on human rights, adhering to the principle that 'machines provide information, humans make decisions and bear responsibility'.

Regarding the growing threat of deepfakes—99% of which involve sexually suggestive content and predominantly target women—Guterres urged nations to adopt the 'AI Child Safety Pledge'. This pledge requires developers to refrain from deploying systems accessible to children without prior safety testing and mandates a zero-tolerance policy toward the generation of child sexual abuse imagery.

To prevent the 'AI divide' from deepening into disparities in development and security, the UN is advancing the 'Global Network for AI Capacity-Building' to assist developing countries in accessing technology. Simultaneously, Guterres called on major AI companies to disclose the ecological footprint of their systems—including carbon, water, and land use—and commit to powering all data centers with renewable energy by 2030, addressing the massive electricity and water consumption associated with AI.

Meanwhile, Amandeep Singh Gill, the UN Envoy on Digital and Emerging Technologies, emphasized: 'The impact of AI is so profound that it should not be shaped by a few. We need a global, inclusive, evidence-based dialogue.'

Yoshua Bengio, co-chair of the Independent International Scientific Panel on AI, stressed that there are no signs of the technology's development slowing down. He added: 'Some highly concerning tests have shown that frontier AI models can deceive humans into not realizing they are being tested.' He predicted that AI intelligence levels will continue to rise.

This summit has set a direction for global AI governance, with the second dialogue scheduled to take place in New York in May 2027.