US President Donald Trump arrived in China on May 13, kickstarting his globally watched high-stakes two-day visit in the shadow of not-so-easy-to-navigate issues including tech, trade, Taiwan, and possibly Iran.POTUS, who is travelling with not just his top officials but more than a dozen business leaders, including Tim Cook and Elon Musk, was welcomed with pomp and pageantry when he landed in Beijing on Wednesday evening around 5.20 PM (IST). While Trump and Xi Jinping are expected to unveil trade deals that Trump can claim as "big deals" back home, the Chinese president is possibly walking into the room with a trump card – Iran. Trump is expected to encourage Xi to push Iran to reopen the strait and to agree to a peace deal. Xi could possibly use this leverage to get what he wants on an issue important to Beijing: Taiwan.
When the two sides meet for bilateral talks, the cameras will focus on trade numbers, Boeing orders, and rare earth commitments. What the cameras won't show is the more dangerous contest unfolding in the background — a full-scale technology war over artificial intelligence that neither side can afford to lose.
Also Read:What Is Thucydides Trap? Xi Jinping’s Message To Trump Explained
This is not a future threat. It is already happening, and the US's own agencies have the evidence.
Just three weeks before the summit, the White House Office of Science and Technology Policy released a formal memo with a striking accusation. OSTP Director Michael Kratsios wrote that "foreign entities, principally based in China, are engaged in deliberate, industrial-scale campaigns to “distil” US frontier AI systems — leveraging tens of thousands of proxy accounts to evade detection and using jailbreaking techniques to expose proprietary information. The White House called it theft. Systematic, coordinated, industrial-scale theft.
That memo landed three weeks before Trump's Beijing flight. Its timing was not accidental.
Most people have never heard the word used this way. Distillation does not require hacking servers or stealing code. Instead, you flood an advanced AI system with millions of carefully designed questions through fake accounts, collect the answers, and use them to train a cheaper rival model that mimics the original. No break-in. No stolen hard drive.
The result: companies that have spent billions and years building the world's most powerful AI systems are having their capabilities quietly copied — without a single line of code being stolen in the traditional sense.
Anthropic, as per several media reports, had already documented the evidence in February — three Chinese AI firms, DeepSeek, Moonshot AI, and MiniMax, had collectively generated over 16 million exchanges with Claude through roughly 24,000 fraudulent accounts. OpenAI reported similar theft of its ChatGPT models. By April, OpenAI, Anthropic, and Google, otherwise fierce commercial rivals, began sharing attack intelligence with each other to fight back.
The White House memo also warned that these distillation campaigns allow foreign actors to "deliberately strip security protocols from the resulting models and undo mechanisms that ensure those AI models are ideologically neutral and truth-seeking". In plain language: China is not just stealing America's AI capability. It may be rebuilding that capability without the guardrails.
Source: India Latest News, Breaking News Today, Top News Headlines | Times Now
