UN puts artificial intelligence governance under the spotlight as global risks intensify

Artificial intelligence is moving higher on the international diplomatic agenda as governments confront a difficult question: how can increasingly powerful systems be developed without creating risks that extend beyond national borders?

The issue is gaining prominence as the United Nations opens its 81st General Assembly session in New York. The UN has already placed multilateral AI governance among its priorities, while Secretary-General António Guterres has called for stronger international oversight of rapidly advancing technologies.

The debate comes amid a widening disagreement over how quickly governments should intervene. In the United States, the administration of President Donald Trump has emphasized technological competition and warned that excessive regulation could weaken the country's position relative to China. Other US lawmakers and technology leaders are calling for additional safeguards.

AI governance moves deeper into the UN agenda

The United Nations is not starting the discussion from scratch. The organization has been working on international AI governance for several years, including through the Global Digital Compact and the Global Dialogue on AI Governance.

At a September 10 diplomatic meeting in New York, UN General Assembly President Khalilur Rahman argued that the international community needs to remain actively involved as AI capabilities expand. He pointed to the technology's potential impact on employment, energy consumption, development and human security.

The UN's approach is broader than a debate over whether individual AI systems are safe. It also encompasses questions of access, international cooperation, accountability and the risk of fragmented regulatory systems.

The Security Council has previously discussed AI in the context of international peace and security, including possible consequences involving cyber operations, disinformation and autonomous weapons.

Washington resists calls for a broad slowdown

The debate in Washington illustrates the difficulty of establishing a common international position.

Trump has rejected calls for a major shift toward tighter federal oversight, arguing that the United States already possesses tools to police the industry. He has also warned that slowing American AI development could benefit China. On September 14, he dismissed growing concerns about AI safety and described the campaign for additional restrictions as a “sick conspiracy.”

House Speaker Mike Johnson has expressed a similar concern about the geopolitical consequences of slowing development. He has argued that a moratorium could give China an advantage and has favored safeguards that do not prevent companies from continuing to innovate.

The argument reflects a central tension in the US debate: policymakers are being asked to address potentially serious technological risks while maintaining American leadership in a strategic competition with China.

Bipartisan pressure for stronger safeguards

The US political debate is nevertheless more complex than a simple division between supporters and opponents of regulation.

Senators from both parties are negotiating legislation that could impose additional obligations on developers of highly capable AI systems. One proposal under discussion would require companies to demonstrate that they are taking measures to prevent major risks, while giving federal authorities greater powers to assess advanced models.

The discussions have intensified following warnings from researchers and technology executives about the possibility that increasingly autonomous AI systems could be misused or become difficult to control.

OpenAI has also backed bipartisan legislation targeting the potential use of AI in the development of biological weapons and synthetic viruses, demonstrating that some regulatory initiatives are focused on specific security threats rather than AI development as a whole.

A global debate over the meaning of AI risk

The concept of an “existential risk” from artificial intelligence remains contested. Some researchers and technology executives argue that increasingly capable systems could eventually create dangers that existing institutions are poorly equipped to manage. Others emphasize more immediate and measurable problems, including cyberattacks, fraud, disinformation, labor disruption and the misuse of AI-enabled tools.

That distinction matters for policymakers because different risks can require very different responses.

China, for example, is developing its own regulatory framework while continuing to pursue rapid adoption of AI. Recent measures have focused on areas such as safety assessments, testing, data security and the behavior of autonomous AI agents.

The European Union is also pursuing a regulatory approach. European Commission President Ursula von der Leyen has recently called for discussions with leading AI laboratories over the risks posed by frontier systems, while emphasizing evaluation and security measures.

The challenge for international institutions

The growing number of national approaches raises a broader question for the United Nations: whether governments can agree on common principles while maintaining different economic, technological and security priorities.

The UN General Assembly's current work on AI governance seeks to provide a forum for that discussion. Its president has stressed that every member state should have a voice in determining how the technology develops and how its benefits and risks are distributed.

For the Security Council, the challenge is narrower but potentially more consequential. Its mandate concerns international peace and security, making questions such as autonomous weapons, cyber operations, critical infrastructure and AI-enabled security threats particularly relevant.

The Council's published September programme does not currently list a dedicated AI meeting, although UN officials have indicated that AI governance and security risks remain part of the wider international discussion.

From technological race to governance challenge

The emerging debate is therefore no longer simply about whether AI should advance. Governments are increasingly discussing how advanced systems should be evaluated, who should be responsible when they cause harm and what forms of international cooperation are needed.

The United States and China remain central to the technological competition, while Europe and the United Nations are seeking different mechanisms for managing the risks. At the same time, AI companies themselves are divided over whether stronger external oversight or industry-led safety measures should take priority.

As the General Assembly's high-level meetings begin in New York, AI is consequently becoming both a technology policy issue and a question of international governance. The central challenge for governments will be to reconcile innovation, economic competitiveness and security without allowing national approaches to create gaps in oversight as AI capabilities continue to evolve.