OpenAI, Google AI sales to blacklisted Chinese firms expose loopholes in US tech curbs

OpenAI and Google have been providing advanced artificial intelligence (AI) models to Singapore-based subsidiaries of Chinese technology giants whose parent companies are on the Pentagon's blacklist, exposing a significant gap in Washington's efforts to curb Beijing's access to cutting-edge AI.

The disclosure, first reported by the Financial Times, has reignited a policy debate over whether the US is focusing too heavily on restricting AI chips while leaving frontier AI software subject to weaker export controls.

According to reports, OpenAI and Google supplied access to their advanced AI models to Singapore-based subsidiaries of Chinese technology giants Alibaba, Baidu and Tencent, companies whose parent entities appear on the US Department of Defence's list of firms allegedly linked to China's military. Although the transactions did not violate existing US laws, they have reignited concerns that export rules have failed to keep pace with the rapid evolution of AI services.

More from Tech

Hardware restrictions are outpacing AI software rules

The latest disclosures underline a key weakness in Washington's AI strategy. US policymakers have invested significant effort in limiting China's access to advanced chips from companies such as Nvidia, recognising that computing power is essential for training frontier AI models. However, commercial access to AI models through cloud-based application programming interfaces (APIs) remains subject to comparatively lighter regulation.

Because the Chinese companies reportedly accessed AI services through overseas subsidiaries registered in Singapore, the arrangements fell within existing compliance frameworks. Yet security experts argue that such structures can still provide sophisticated AI capabilities to organisations that Washington seeks to limit.

The issue has become even more sensitive amid concerns over "model distillation", a process in which outputs from advanced AI systems are used to train competing models. OpenAI confirmed that it recently suspended API access for some Alibaba-affiliated users after identifying concerns related to potential misuse. Google, meanwhile, maintains that its policies prohibit harmful or prohibited use of its AI systems, while acknowledging that enforcing such restrictions globally remains challenging.

AI becomes the next battleground in US-China competition

The controversy reflects a broader shift in geopolitical thinking: AI models themselves are increasingly being viewed as strategic assets rather than merely software products. The debate has moved beyond who manufactures the world's most powerful chips to who controls access to the intelligence running on them.

The disclosure comes as both Washington and Beijing are expanding controls over advanced AI technologies. The United States has tightened scrutiny of frontier AI models alongside semiconductor exports, while China is reportedly considering restrictions on overseas access to its own advanced AI systems, signalling an emerging era of reciprocal technology controls.

The differing approaches among American AI companies also illustrate the absence of a uniform industry standard. While OpenAI and Google have permitted limited overseas access under compliance safeguards, Anthropic has adopted a far stricter policy by blocking all Chinese entities from using its frontier models, citing security concerns.

For policymakers, the episode raises a larger question: if AI models increasingly determine economic competitiveness and national security, should governments regulate software exports with the same intensity as semiconductor shipments? As AI becomes central to military, industrial and scientific capabilities, the answer is likely to shape the next phase of the global technology rivalry.