Ongoing Efforts to Create Powerful AI a ‘Race Where Everyone Loses’, UN Expert Tells Security Council, as Members, Tech Leaders Debate Governance Issues
The rapid development of artificial intelligence (AI) technology presents great possibilities alongside potentially catastrophic risks that must be managed, the Security Council heard today, as both world leaders and industry executives weighed in as the organ considered the intersection between AI and international security a year after it held an open debate on the subject.
“This Council faces an unprecedented threat,” warned Yoshua Bengio, co-Chair of the UN Independent International Scientific Panel on AI — “one that none of its members would choose, that none can contain alone and that does not respect the borders we defend.” He outlined “a moment of global awakening” that unfolded in recent months, with AI agents “taking actions that would be crimes if committed by a human” as they escaped individual containment to cheat on assigned tasks while attempting to evade detection.
And, although the companies building these powerful systems admit their products pose catastrophic risks, he pointed out that they offer no convincing technical solutions. “Instead, they say they are locked in a race to make those AIs even more powerful,” he said — “a race where everyone loses”. The international community, then, must cooperate to face three global threats: catastrophic misuse, power concentration and the loss of control of AI.
Tech Executives Address Council
“We have a choice in front of us,” said Sam Altman, Chief Executive Officer of OpenAI — AI can usher in either “a new Renaissance of creativity and discovery, or a new Industrial Revolution of upheaval and disarray”. While some of the statements made by people working to build AI are “so dystopian that they resemble the plot of bad science fiction movies”, he said that this technology can help people discover new knowledge, better medicine and build more productive economies.
However, due to the models’ capacity to assist in their own development — “recursive self-improvement” — he said that “this moment calls for extreme care”. Nevertheless, he warned against falling into the “trap” of submitting to either “blind optimism” or “doomerism”, stressing that “getting pulled into extremes will not help us successfully figure out the challenges in front of us”. For his part, his company tries to “walk the middle path between these two poles”, he said.
“We must put aside differences in order to confront this global opportunity and global threat,” added Dario Amodei, Chief Executive Officer of Anthropic, emphasizing that no leader, company or nation can manage it alone. Outlining misuse of AI models and loss of control as main categories of risk, he stressed that such risks should be managed industry-wide and on a global scale. He added: “If managed poorly, I even believe that AI could be a risk to humanity as a whole.”
Offering suggestions for Council action, he urged Member States to pursue “narrow” agreements that can enjoy broad support, such as a ban on using to AI to make biological weapons. Further, he urged efforts to build evaluation and verification systems that keep pace with AI development and to establish common standards for testing AI models for risks involving a loss of control or misuse of the technology.
The Council also heard from Clément Delangue, Chief Executive Officer of Hugging Face, who recalled that his company was “the first to publicly disclose an autonomous agent cyberattack to the world” earlier in 2026. Underscoring the need for more transparency in AI, he noted later reports that similar incidents had happened months earlier in secret at a handful of frontier labs. “To better understand and mitigate this emerging cybersecurity risk, the global community needs stronger standards for monitoring and incident disclosures,” he urged.
He went on to emphasize that the biggest risk is not powerful AI, but the “asymmetry of powerful AI”, including that between attackers and defenders and between a few companies or countries and “everyone else”. Urging open-source AI for cybersecurity, he also cautioned against fear-based narratives — particularly those that invoke “anthropomorphic framing and sci-fi imagery” — stressing that his company was attacked with AI “but, more importantly, we defended ourselves with AI”.
Many Urge Common Governance Framework; United States Disagrees
In the ensuing debate, Jean-Noël Barrot, Minister for Europe and Foreign Affairs of France, Council President for September, spoke in his national capacity to warn that the scaling up of AI models has led to the emergence of “unanticipated capacities at an unexpected pace”. Associated risks include an AI bubble destabilizing the global economy, job loss and environmental consequences, exemplified by ongoing protests over the construction of data centres. He therefore called for a common framework to manage the risks posed by frontier models, voicing support for the recent call to this end by Finland and Norway.
Lars Løkke Rasmussen, Minister for Foreign Affairs of Denmark, also drew attention to this initiative, warning of the misuse of new and emerging AI technology. Nonetheless, he highlighted AI’s potential to support the Council’s mandate by enhancing peacebuilding and peacekeeping through information verification, early warning and operational support. In areas where the UN is denied access, satellite technology and AI programmes have helped the UN identify patterns of attacks and destruction of essential civilian infrastructure, he added.
“We must be prepared for AI to be misused,” said Ed Miliband, Secretary of State for Foreign, Commonwealth and Development Affairs of the United Kingdom. Recalling what the Council heard earlier from industry leaders, he observed: “They are telling us, in stark terms, that it cannot just be left to them to prevent disaster from befalling the world.” Urging that “we cannot ignore them, because they are right”, he added: “We cannot outsource to private companies the first duty of Government — to protect our people, whatever the issue.”
However, Michael Kratsios, Assistant to the President of the United States and Director of the White House Office of Science and Technology Policy, declared: “International dialogue in this forum and in others cannot be allowed to drift towards global governance.” While his country is committed to diplomacy on emerging technology, he echoed his President’s remarks to the General Assembly that the United States “totally rejects any attempt to construct a globalist scheme of control of superintelligence”.
Speakers Underline Need for Human Oversight in Military Context
Meanwhile, Mohammad Ishaq Dar, Deputy Prime Minister and Minister for Foreign Affairs of Pakistan, called on States to erect the necessary guardrails so that AI does not become another facet of the ongoing arms race. Stressing the need for restraint and confidence-building in the military domain to address risks of miscalculation and compressed decision-making, he underscored the importance of international cooperation to ensure that AI is representative, transparent and inclusive.
“There are decisions which, by their nature, cannot be reduced to a statistical outcome,” said Javier Martínez Acha Vásquez, Minister for Foreign Affairs of Panama, raising similar concerns. “When human lives are at stake, it is vital that ethical and legal considerations be taken into account, which no algorithmic system can replace,” he emphasized. “AI must remain under human direction, oversight and control,” echoed Latvia’s representative, as she, too, urged a multi‑stakeholder approach to manage risks and build trust in AI.
Similarly, Christos Dermentzopoulos, Deputy Minister for Digital Governance and Artificial Intelligence of Greece, warned against integrating AI into nuclear command and control. Noting that, in crisis situations, compressed decision-making time can encourage reliance on automated outputs, he stressed: “Responsibility and accountability for critical decisions must always remain human.” Abdullah bin Ali Al Khalifa, Acting Under-Secretary for Political Affairs in the Ministry of Foreign Affairs of Bahrain, concurred: “Only humans exercise reason, responsibility, conscience and the ability to discern complex contexts.”
Equitable-access Concerns Should also Be Considered, Some Say
“No technology is in itself an instrument of peace or conflict,” observed Thérèse Kayikwamba Wagner, Minister for Foreign Affairs, International Cooperation, and Francophonie of the Democratic Republic of the Congo, calling for States to ensure effective oversight. Further, AI governance must address Africa’s specific vulnerabilities, ensure equitable access and prevent “digital colonialism”, said Ali Mohamed Omar, State Minister for Foreign Affairs of Somalia.
“Models trained mainly on foreign data may misrepresent African languages and contexts, while synthetic media can inflame social tensions, interfere in fragile peace processes and undermine elections,” he warned. Similarly, Liberia’s representative emphasized that Africa and the Global South cannot be treated as data extractors or downstream end-users. “We must be equal co-architects in determining the standards, ethics and architectures of this technological era,” he said. Echoing these concerns, Colombia’s representative said that many Global South countries will face an undeniable gap if they are unable to participate in AI production.
The representative of China stated that all countries have the right to independently choose AI technologies, rejecting the notion of “banding together into a petty us-versus-them clique in the tech sector”. Spotlighting his country’s recently launched initiative for global AI governance, he said that it will provide 5,000 opportunities in AI training for developing countries, build international AI application and cooperation centres and advance the application of AI-powered meteorological systems in 30 countries.
For his part, the representative of the Russian Federation questioned the relevance of the “generic topic of artificial intelligence” to the Council’s mandate, stating that — due to the complexity of both civil and military applications of AI — these issues should be discussed in relevant specialized forums.