Global rules 'urgent': AI 'Godfather' pushes against Trump's rejection of guardrails

US President Donald Trump dismissed concerns about artificial intelligence (AI) as a hoax during a speech at the United Nations General Assembly (UNGA). Professor Yoshua Bengio, one of the "Godfathers" of AI and co-chair of the UN's Scientific Panel on AI, strongly disagreed with the President's stance. Professor Bengio will address the UN Security Council to advocate for urgent action on AI safeguards.

Bengio emphasized that big tech companies are taking significant risks with their AI development and that decisions should not solely be made by a few CEOs or governments. He stressed the importance of international agreements on safeguards to ensure human control and preserve human dignity. The Turing Award-winning professor highlighted that AI development should remain within human control, as uncontrolled AI could lead to disastrous consequences.

While acknowledging the US's rejection of a joint statement from 22 countries, including Australia, urging humans to remain in control of AI, Professor Bengio appreciated the effort. He called for continued international cooperation and urged more countries to join this cause. Bengio believes that when nations form coalitions and pool their resources, they can shape a future aligned with their values and maintain sovereignty in a multilateral world.

The US-China AI rivalry, with American companies reportedly six months ahead of Chinese rivals, does not negate the need for cooperation, Bengio argued. He stressed that both nations are rational actors with the same desire for a good life for their citizens. Bengio believes that both the US and China have an interest in sitting down and agreeing on principles to guide AI development.

Professor Bengio warns that the most significant AI risks arise from recursive self-improvement, where advanced AI systems autonomously design even more powerful and uncontrollable successors. He urged world leaders to act urgently to prevent losing control over AI, as such outcomes would affect all humanity.