Secretary of State Marco Rubio confirmed on Tuesday that the United States and China have held technical-level discussions on artificial intelligence safeguards but have yet to reach any formal agreement, describing the talks as 'still a work in progress' ahead of key multilateral summits. The remarks came after a G7 ministerial meeting and signal that one of the most consequential technology-governance disputes between the world's two largest powers remains unresolved.

What Rubio Said

'I know there's been conversations with them sort of at a — at a technical level, but we don't have anything to announce on it,' Rubio told reporters. He added: 'Obviously, we've heard their perspective, they've heard ours, but that's still a work in progress.' The comments came in response to questions about whether Washington and Beijing had held substantive discussions on AI guardrails since their last presidential-level meeting.

Rubio was explicit that the two governments had each laid out their positions without converging on a shared framework. He declined to characterise the talks as a breakthrough, instead framing them as preliminary exchanges at the working level.

Multilateral Forums in Focus

Rather than limiting the AI safety debate to a bilateral channel, Rubio indicated that the issue would be elevated to broader international platforms. 'It should be an issue that we raise at the summit,' he said, referencing the forthcoming Asia-Pacific Economic Cooperation (APEC) gathering, where he expected AI governance to feature prominently. The subject is also expected to be on the agenda at the Group of 20 (G20), where China holds a seat.

This comes amid a wider global push to establish norms around advanced AI systems, with the United Kingdom, the European Union, and several Asian economies having already convened their own AI safety forums. The absence of a binding US-China framework is widely seen as a gap in the emerging international architecture.

America-First Framing on AI Policy

Rubio linked Washington's approach to AI governance to a broader 'America First' foreign policy posture. 'The Chinese system is always going to act in the best interest of the Chinese system, and the United States will always act in the best interests of the United States,' he said. He acknowledged that this alignment of national interests would at times put the two countries in direct conflict.

'We don't work for the global order. We don't work for other countries. We work for the United States,' Rubio said, underscoring that the Trump administration would measure its positions on AI — and other issues — by what it deemed beneficial to Americans, even when that created disagreements with Beijing.

Prospect of US-Russia-China Nuclear Talks

Separately, Rubio addressed questions about a potential trilateral meeting involving the United States, Russia, and China on nuclear weapons and proliferation. He said there was an intention to hold such a meeting and that officials were working on its organisation, but offered no date or formal announcement. He noted that President Donald Trump wanted the meeting to take place if possible and viewed the global status of nuclear weapons as a subject worth addressing at the highest levels.

Rubio declined to speculate on what China might be willing to accept in such a format, keeping expectations deliberately low-key.

What to Watch

With APEC and G20 summits approaching, AI governance is set to become a live diplomatic flashpoint. Whether Washington and Beijing can move beyond technical-level exchanges to a concrete joint statement — or even a bilateral mechanism — will be a key indicator of how much room exists for US-China cooperation in an otherwise strained relationship. All eyes will be on whether the two sides use the multilateral platform to narrow differences or harden positions.