Altman, Amodei urge UN Security Council to strengthen AI safety controls

MUMBAI: OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei have urged the UN Security Council to strengthen safeguards around artificial intelligence, warning that increasingly capable and autonomous AI systems could outpace institutions and weaken human oversight.

Speaking at a Security Council briefing on Wednesday, Altman highlighted the dual potential of AI, saying the technology could accelerate scientific discovery, healthcare, education and economic productivity while also creating risks if systems become too autonomous.

“AI systems become more capable and more autonomous, they could move faster than our institutions, concentrate power in too few hands, or make decisions that people no longer understand or control,” Altman told the Council.

Altman reiterated his call for frontier AI developers to monitor the pace of development and avoid taking excessive technological risks in the race to build more capable systems.
He said the industry must avoid situations where AI progresses so rapidly that people can no longer understand what is happening or intervene when necessary.

According to Altman, competition between companies should not become a reason to make what he described as rash decisions around AI development. He also said frontier labs should avoid building systems without adequate alignment, monitoring and safety guarantees.

Altman further stressed that people should remain at the centre of AI decision-making, arguing that no individual company or country should be able to use powerful AI models to impose its worldview on others.

Amodei called for international cooperation on AI regulation, proposing a series of measures focused on global security.
He suggested beginning with narrow international agreements, including commitments against using AI to develop biological weapons or allowing AI technologies to facilitate their development.

The Anthropic CEO also called for evaluation and verification systems that can keep pace with frontier AI development. Such systems, he said, could give governments greater visibility into the capabilities of advanced models and help countries verify their commitments.

Amodei additionally proposed common global standards for testing AI models against loss-of-control and misuse risks, along with a notification system for significant AI incidents affecting global security.

The comments from the two AI executives underline the growing focus on how governments and technology companies should manage increasingly capable AI systems.
Both executives pointed to the need for human oversight while arguing that the benefits of AI should continue to be developed responsibly. Their remarks at the Security Council also put AI safety, model evaluation and international coordination into the broader discussion on global security.

While neither executive endorsed an extinction-level scenario, both called for stronger safeguards and international cooperation as AI systems become more capable and autonomous.