UN and Vatican push for global AI governance amid US-China race

By Splitfeed EditorialLast updated Sep 26, 2026, 20:18 UTC

On September 21, 2026, reporting from the New York Times, Nikkei Asia, and Religion News Service converged on a common thread: the US-China artificial intelligence competition is intensifying, with self-improving AI models emerging as a significant marker of that acceleration. On the same day, the United Nations was reported to be actively seeking a governance role in AI development, while the Vatican, covered through the weekend into September 23, pressed separately for international safeguards. These are distinct institutional efforts, one from the primary multilateral political body and one from a global religious authority, both responding to a race being run almost entirely by two national powers.

The core tension here is whether meaningful governance can be built around a competition that neither dominant participant has shown much interest in slowing. Self-improving AI models, systems capable of enhancing their own capabilities without direct human intervention at each step, represent a qualitative shift in what the race involves. If such models are genuinely advancing, the window for establishing effective international norms may be narrowing, which is precisely the argument both the UN and the Vatican appear to be making, though from different foundations of authority. What remains genuinely disputed is whether either institution has the leverage to influence the two leading states, and whether governance frameworks proposed outside the US-China dynamic can carry any practical weight.

With no published reactions yet from the Pro-AI, Anti-AI, or Middle Ground camps on this particular story, their likely positions follow from their broader patterns. Pro-AI voices would typically argue that competition drives faster breakthroughs and that governance attempts by bodies like the UN risk slowing beneficial progress without meaningfully constraining bad actors. Anti-AI voices would be expected to treat self-improving models as precisely the kind of development that validates calls for urgent international oversight, and to criticize both the US and China for prioritizing strategic advantage over safety. The Middle Ground camp would likely welcome governance conversations while questioning whether the UN or Vatican have the enforcement mechanisms to make their appeals stick.

The argument will sharpen around concrete governance proposals from the UN, particularly any formal mechanisms or binding frameworks it puts forward, and around further technical reporting on what self-improving models can actually do as of late 2026. Either development would give all three camps much more specific ground to fight over.