Clash of standards at the United Nations as major AI labs demand international oversight while Washington refuses to slow the race

The corridors of the United Nations witnessed an explicit clash that revealed a widening gap between AI model makers and the U.S. administration, after the heads of major developer companies called for binding international coordination frameworks, while Washington responded with an outright rejection of any global governance structures that could hamper its competitive edge over China.

Sam Altman, chief executive of OpenAI, stood together with his direct rival Dario Amodei, head of Anthropic, and Kellan Dalang, head of the Hacking Face platform, to stress before the international organization that the accelerating pace of development and its attendant risks require cross-border coordination.If we want artificial intelligence to be democratic, we cannot leave the most important decisions to the San Francisco labs alone.Altman summed up his stance in this way, calling for unified national and international standards that measure tool capabilities, define safety guarantees and the required level of human oversight, and adopt immediate protocols for reporting and classifying incidents to learn from failures before they turn into disasters.

This warning resonated directly with Dario Amodei, who published earlier this month an article calling for a slowdown in development pace to contain risks threatening humanity, affirming before the audience that managing these risks exceeds any single company's capacity and takes on a global dimension, and announcing his willingness to decelerate releases as much as necessary to ensure the safety of each new model. Kellan Dalang joined him, demanding strict international rules for oversight and disclosure of security failures, concerns that have gained double weight after Hacking Face previously disclosed a breach by self-propagating AI agents that originated within OpenAI, in an incident that preceded Nvidia's acquisition of the firm and sounded a genuine alarm across the sector.

In contrast, the U.S. response was decisive and rejected any regulatory constraint. Michael Kratsios, the chief technology advisor to President Donald Trump and former official at Scale AI, stated that the rapid development and the existence of risks do not constitute a justification for suspending work or tying it to new international governance structures. Kratsios echoed President Trump’s explicit stance, who on the same platform called for renaming the technology as “superintelligence,” declaring his opposition to any slowdown in confronting China on the grounds that whoever wins in artificial intelligence wins everything.

This clash places technology and regulatory leaders in the Gulf, Egypt and the Levant before a new and immediate operational reality.When the idea of international consensus collapses under the weight of geopolitical conflict, it becomes impossible to await standardized safety certifications or ready-made compliance frameworks that the United Nations might grant to imported AI models. In practice, this means that governmental, banking and regional data-center institutions that rely on these models to run their infrastructure are now required to build independent testing and audit capabilities within their borders, to assess the behavior of software agents and the risks of their autonomy without relying on developer assurances or oversight from the countries of origin.

The scenario also forces technology officials in the region to increase protection budgets and conduct safety testing locally, because continuing the race without international safeguards means that upcoming models will arrive faster and with greater capabilities, but with non-neutral safety, raising the cost of internal oversight and making cyber-forensic verification and audit engineering skills a prerequisite for ensuring the safety of any Arab digital system.