kimberly-sutanto
US President Donald Trump met with dozens of technology executives at the White House this week for a wide-ranging discussion of artificial intelligence, opening the door to a new phase in the administration's approach to the technology — one that pairs aggressive promotion with explicit reluctance to extend the hand of cooperation to China.
The gathering took place ahead of what has been billed as a "superintelligence" luncheon at the White House, a meeting that brought together the chief executives of some of the world's most influential technology companies. For an industry accustomed to being regulated, scrutinised and occasionally penalised by governments on both sides of the Atlantic, a sitting president hosting the sector's most powerful figures is itself a statement of posture.
Ahead of the luncheon, Trump unveiled a new AI-powered website for government services, a move presented as a demonstration of what the technology can deliver when it is allowed to move quickly. Accompanying the launch, the president vowed not to stifle the technology's growth — a signal aimed at an industry that has spent much of the past year bracing for tighter rules, litigation and political backlash.
The platform is intended to show the technology in an applied, citizen-facing setting, rather than as an abstract experiment. Government portals built around automated assistance and AI-driven tools have become an increasingly common test case for whether public institutions can absorb the technology without surrendering the accuracy, privacy and due process that public services demand. Launching one from the White House elevates the experiment well beyond an ordinary agency pilot.
Analysts note that the timing carries its own message. By pairing a flagship public-sector deployment with an anti-stifling pledge, the administration is presenting AI as infrastructure to be accelerated rather than a hazard to be contained. The framing is deliberately at odds with the more precautionary climate that has dominated policy debates in Europe and in parts of the US bureaucracy.
At the centre of the discussions, however, was a clear line on geopolitics. Trump said he did not want to work with Chinese President Xi Jinping on AI oversight, closing off the possibility of joint talks on safety standards. The remark underlines how thoroughly AI safety has been recast as a matter of national competition rather than shared international governance.
The US-China technology rivalry has for years centred on chips, export controls and market access. Extending the same logic to model safety talks is a significant escalation of the frame. Co-operation on nuclear weapons testing, pandemic preparedness or aviation standards once provided narrow channels for the two governments to engage; efforts to coordinate frontier model rules would cut straight through the competitive core of the technology sector, where Washington's view of Chinese firms is dominated by security suspicion.
Rejecting such a channel also removes what many outside observers see as a potential point of contact. Some researchers have argued that the largest risks in advanced AI are global in nature — capable of crossing borders regardless of the policies of any single government — and that safety norms, if they are ever to be credible, will need to be negotiated between rival powers. Washington's position, as articulated at the summit, is that the competition itself takes priority.
The summit also landed against a turbulent backdrop elsewhere in the industry. Separate reporting from France 24 this week noted that Trump said AI companies have signed a voluntary accord on safety controls — a framework built on commitments rather than statutory obligations. The voluntary model has become the pivot point of an unresolved debate: for industry, it offers speed; for critics, it offers no enforcement mechanism and no remedy when commitments are quietly dropped.
The contrast with events in the private sector could hardly be sharper. OpenAI cancelled the release of a new artificial intelligence model over safety concerns, a decision that suggests at least some developers still consider a voluntary framework insufficient when the stakes rise. Accelerating deployment in government while a leading lab pulls a model from public release sends contradictory signals about how much assurance the technology can absorb before meaningful harm is possible.
That tension is likely to define the period ahead. The administration's posture is unambiguous in its economic direction: keep the United States competitive, keep regulation light, and keep the government from becoming an obstacle. What remains unresolved is whether the institutions in Washington can build oversight frameworks that are flexible enough to satisfy the industry and credible enough to satisfy the public, and whether partners abroad will accept American safety norms that are themselves shaped by a commitment to exclude China.
The economic stakes are considerable. Companies that can point to real-world deployments, including public-sector ones, gain credibility with customers, investors and regulators. Governments that move early on AI services risk reputational damage if automated tools produce errors at scale, but risk political failure if citizens perceive that basic services have been handed over to systems that cannot explain their decisions.
Not every item on the agenda was strictly business in nature. The same programme also turned to a story from South Korea, where dozens of dogs were rescued from a breeding farm as the country prepares for a complete phaseout of the dog meat industry. The rescues are part of a longer transition process, with animal welfare organisations and campaigners continuing to press for a full end to the trade.
South Korea's move has drawn sustained international attention, and the associated coverage tends to attract both strong moral support and resistance from stakeholders who argue that the transition must not simply displace the industry without alternatives for the farmers and traders involved. The phaseout is being watched as a test of whether regulatory intent in a global economy can be matched by practical support on the ground.
Pairing the two stories in a single business segment is itself revealing. A US president promising to liberate AI from regulatory constraint, a South Korean government moving to extinguish an industry by law, and a lab withdrawing a product on safety grounds all belong to the same wider debate about who should decide what the technology and the market are allowed to do — and how fast.
For now, the administration's answer is clear on speed. The question for the weeks ahead is whether the voluntary accord announced this week will hold, whether the government portal delivers, and whether any route to a US-China conversation on AI safety survives a summit designed to make such a conversation politically impossible.
Comments (0)