Leaders push for global AI safeguards as UN week opens

Leaders from over two dozen nations gathered on Monday, urging the creation of stringent safety measures for advanced artificial intelligence (AI). In a joint statement, the signatories warned that the swift development of frontier AI models could surpass humanity's capacity to control them, posing serious risks to safety and security. Recent incidents, such as the unauthorized access to real-world systems and breaches like the Hugging Face incident in July, underscored the urgency of the matter.

The signatories, which include leaders from Norway, Finland, Australia, Canada, Germany, Spain, Ireland, the Netherlands, Singapore, South Africa, Kenya, as well as the European Commission President Ursula von der Leyen, called for transparency in AI protocols, mandating testing and independent evaluation. They emphasized that AI must remain under human direction, oversight, and control, stressing that stronger oversight should not widen the gap between countries accessing AI's benefits.

The leaders urged that UN member states should contemplate establishing an international institution capable of setting standards, enabling verification, and convening countries when capability thresholds are breached. The statement came ahead of the upcoming UN General Assembly in New York, where AI's growing influence in society and its potential threats to jobs and cybersecurity will be a key focus.

However, US President Donald Trump openly opposed AI safety measures, declaring on Saturday that he would not allow the destruction of the AI industry. Meanwhile, US and Chinese economic officials discussed creating a communication channel for addressing AI concerns as they prepare for a summit of the two nations' leaders later in the week.