AI regulation turns into a global power struggle—UN warnings, EU rules, and US-China AI control

A cluster of UN and policy interventions on 24–25 September 2026 is converging on one theme: advanced AI is moving faster than governance, and governments are reacting in real time. Bangladesh’s prime minister used a maiden UN General Assembly address to call for international action on the Rohingya crisis and to frame Bangladesh’s post-2024 “uprising” rebuilding around a “3-R” strategy, while also signaling support for Palestine. Thailand’s prime minister, also speaking at the UN, pushed for stronger protection of Palestinians and renewed dialogue with Cambodia, as Fiji warned that an HIV epidemic could worsen without climate finance and UN reform. In parallel, Australia’s prime minister warned in a UN speech about the “furious pace” of AI after a security breach, while the Pope and European lawmakers debated how the EU should answer frontier AI risks and provider responsibility. Strategically, the common thread is that AI governance is becoming an arena for geopolitical leverage, not just consumer protection. The EU’s emerging “EU Kids Act” approach—protecting minors from addictive platform designs—faces criticism that age-based rules could deepen exclusion for disabled users, migrants, and the undocumented, turning regulation into a legitimacy test. Meanwhile, Xi Jinping is urging limits on US-China rivalry and calling for “human control of AI,” a signal that Beijing wants guardrails that preserve its strategic autonomy while constraining adversary advantages. The Council on Foreign Relations frames the challenge as how to reach an AI arms-control deal with China, implying that verification, definitions of “agentic” systems, and incident response will be central bargaining chips. Even the information environment is part of the contest: experts say AI-generated protest videos on TikTok are designed to amplify polarization, raising the stakes for cyber and influence operations. Market and economic implications are likely to concentrate in AI compliance, cybersecurity, and platform governance spending. Australia’s government breach narrative and the “agentic AI” question around the Medicare breach point toward higher demand for identity, endpoint, and incident-response tooling, which can lift risk premia for insurers and government contractors tied to critical-infrastructure security. EU lawmakers’ focus on provider responsibility and minors’ protections suggests near-term regulatory-driven costs for social platforms, ad-tech, and model providers, with potential knock-on effects for ad targeting revenues and app monetization strategies. On the macro side, UN speeches linking climate finance, public health, and UN reform reinforce that sovereign and development finance flows may be re-priced, especially for countries exposed to climate and health shocks. In the background, the US-China AI rivalry debate and “human control” messaging can influence export-control expectations and investment decisions in frontier compute, semiconductors, and cloud services, even if no single commodity shock is directly cited. What to watch next is whether these speeches translate into enforceable timelines, technical standards, and cross-border incident norms. Key indicators include EU legislative movement on frontier AI provider liability, implementation details of the EU Kids Act, and any follow-on guidance after Australia’s security breach warning—particularly around “agentic” system oversight and auditability. On the diplomacy track, monitor signals from UN reform negotiations and any concrete steps toward renewed dialogue between Thailand and Cambodia, because humanitarian protection and institutional credibility are being bundled together. For great-power bargaining, track whether Xi’s “human control” framing is echoed in US and EU proposals for AI arms control, including definitions, verification mechanisms, and incident reporting thresholds. Finally, watch for escalation in the information domain—such as platform enforcement actions against AI-generated political content—and for measurable changes in cyber threat reporting tied to government and healthcare systems.

Geopolitical Implications

• 01

  AI governance is becoming a bargaining domain for great-power competition, with “human control” and arms-control concepts likely to shape future standards and verification regimes.

• 02

  EU and national regulators are moving from principles to enforceable liability and platform-safety rules, potentially creating regulatory fragmentation that affects cross-border AI deployment.

• 03

  Cyber incidents and healthcare/critical-infrastructure breaches are likely to drive faster adoption of security-by-design requirements and auditability demands.

• 04

  Humanitarian and UN reform agendas are being used to build coalitions, which may influence how states coordinate on AI safety and information integrity.

Key Signals

• —Drafting and voting milestones for EU frontier AI provider responsibility and the EU Kids Act implementation details.
• —Official follow-up from Australia on the scope of the security breach and any new oversight framework for AI systems used in government.
• —US and China alignment or divergence on definitions of “agentic AI,” human control thresholds, and incident reporting norms.
• —Platform enforcement actions against AI-generated political content and any measurable changes in polarization-amplification indicators.