Hong Kong must wake up to the cold hard geopolitics of AI

With AI increasingly being treated as a strategic asset, the city should ensure it has access to the technologies shaping the global finance US President Donald Trump answers questions from reporters after signing an executive order in the Oval Office of the White House on March 31, in Washington, as Commerce Secretary Howard Lutnick listens. Photo: AP Officials talking about artificial intelligence (AI) often speak in terms of the next industrial revolution or existential risks. Such rhetoric ultimately does little to capture how AI filters into the ordinary workings of a city like Hong Kong. The geopolitics of AI are no longer abstract. They are beginning to shape who can use which tools – and on what terms. Over the past few months, two major international banks in the city have reportedly felt the impact. Anthropic’s Claude, a generative AI tool used by staff at JPMorgan and Goldman Sachs, was apparently removed from internal systems used by Hong Kong-based teams. The move amounts to more than an inconvenience. If colleagues in Singapore or New York retain access while those in Hong Kong do not, the city’s position as a global financial hub is weakened. Claude’s removal is not merely a corporate adjustment. It is a warning that AI is increasingly treated as a strategic asset, rather than as neutral infrastructure. Behind those adjustments lies a more complex geography of access. Some advanced systems now include regional carve-outs excluding “Greater China”. Anthropic has made clear that Claude is not officially supported in Hong Kong. For local employees, the reasoning matters less than the effect: tools treated as routine elsewhere can disappear overnight. The situation became clearer in June, when Anthropic received a US export control directive requiring it to suspend foreign nationals’ access to its newest models. The company temporarily disabled the systems worldwide. Pages from the Anthropic website and the company’s logo are displayed on a computer screen in New York on February 26. Photo: AP Weeks later, US authorities eased their position. Anthropic’s flagship model, Mythos 5, was later restored to a set of trusted organisations and the need for export licences was removed. Access resumed under revised conditions. Last week, Chinese President Xi Jinping delivered a keynote address at the World Artificial Intelligence Conference in Shanghai, casting AI as a field for international cooperation. By laying out China’s ambitions, the speech positioned AI as a strategic lever in Beijing’s technological rivalry with Washington as well. Taken together, episodes like these underline something more consequential: the most capable AI systems are no longer distributed like ordinary software. They now sit at the intersection of corporate terms and state strategy. Access can shift abruptly. Hong Kong is in the crosshairs. Washington has signalled that frontier AI is strategically sensitive and that overseas users can be cut off at short notice. This creates a new operational reality: evaluating AI no longer means reviewing only vendor contracts or internal risk controls. It also means monitoring geopolitical risk. President Xi Jinping waves as he arrives at the World AI Conference in Shanghai on July 17. Photo: EPA It also raises a commercial question: if companies conclude that Hong Kong staff cannot reliably use the same systems as colleagues in rival centres, future mandates may shift elsewhere. In the age of strategic AI, competitiveness depends not only on capital and connectivity but also on continued access. What, then, is in Hong Kong’s control? The city has no single AI statute. The Digital Policy Office provides an Ethical AI Framework and a separate guideline on generative AI. Financial regulators have issued policy statements and circulars which financial institutions are expected to follow. Unhandled type: inline-plus-widget {"type":"inline-plus-widget"} This principles-based approach contrasts with the increasingly strategic posture adopted elsewhere. For organisations operating across these systems in Hong Kong, the divergence is stark. Whether a system can be used may hinge on a contractual clause, a licensing term or another jurisdiction’s export restriction. Operational continuity, once taken for granted in cloud software, is no longer assumed. From one angle, these developments are quite familiar to Hong Kong. The city has long been recognised as a place where different systems conduct business – common law and Chinese sovereignty, Western capital and mainland industry. With AI, that role is becoming more complex. It now involves not only legal and financial regimes but also technical standards, data policies and decisions about who is permitted to access the most advanced systems. Recent episodes involving Claude show how quickly that access can change once advanced models are treated as strategic assets. A useful tool can vanish overnight. In that sense, the story of AI’s true power is not just about one company. It reflects how AI, once marketed as generic cloud software, has become part of the machinery through which states and firms negotiate advantages. For Hong Kong, the question is no longer whether AI will shape competitiveness, but whether the city can avoid being treated as a carve-out when access to the most powerful systems is allocated, and whether its policy choices are enough to ensure it remains connected to the technologies shaping global finance. That depends less on rhetoric about innovation and more on whether Hong Kong can exercise regulatory clarity, institutional reliability and geopolitical trust as advanced technologies become increasingly entangled with state strategy.