Leading artificial-intelligence executives warned theUN Security Councilthat increasingly capable systems could create serious security risks if development and deployment move faster than safeguards. UN News reported that OpenAI and Anthropic were among the technology leaders briefing the council on September 23 during the General Assembly period. Speakers described AI’s benefits across education and other sectors, but also emphasized that systems used in sensitive environments need governance, oversight and cooperation. The session was a briefing and warning, not a new binding United Nations regulation or an AI shutdown order.
Anthropic chief executive Dario Amodei told the council that poorly managed AI could become a risk to humanity as a whole, according to Reuters. OpenAI CEO Sam Altman argued that international cooperation is necessary and that important decisions should be shaped through democratic processes and governments accountable to the public. Those remarks placeAI securityalongside technical performance in the global debate: safety questions now include who sets standards, how incidents are reported and how systems are governed across borders.
Hugging Face co-founder Clément Delangue described AI as both an offensive and defensive cybersecurity tool. Reuters reported that he discussed an incident involving AI agents and said his organization used another model to help defend its infrastructure. The example illustrates why the debate is not limited to future speculation. AI can assist defenders while also creating new attack paths, making testing, access controls and human review important. The sources do not establish a new international cybersecurity rule or say that any specific model has been banned.
AI scientist Yoshua Bengio told the Security Council that dangers from systems beyond human control could represent an unprecedented threat that no country can contain alone. UN News linked the discussion to theGlobal Digital Compact, which seeks broader cooperation on digital technology and AI. The practical takeaway is cautious: governments are debating guardrails and accountability while capabilities advance. For Philippine readers, the briefing is international context rather than a local policy change, product launch or confirmation that a new AI rule already applies domestically.
