The Chinese innovation that could give Xi Jinping an 'enormous' edge in warfare that has the Pentagon scrambling

By CHARLIE SPIERING, US POLITICAL REPORTER Published: 13:22, 4 April 2026 | Updated: 13:22, 4 April 2026 The United States has developed and deployed Artificial Intelligence on the battlefield, as demonstrated in the Iran war, but there's one enormous leap in the technology that would threaten the United States. 'AI warfare is here, and it's expanding, and I don't think that that's going away. If anything, I think it's getting bigger and more strong,' Wynton Hall, the author of the New York Times best selling book 'Code Red: The Left, the Right, China, and the Race to Control AI' in an exclusive interview with the Daily Mail. Currently, the United States is using AI to sift through the enormous pool of collected data for useful intelligence, such as images, audio, video, satellite feeds.  The military has used AI tools such as Anthropic's Claude for real-time targeting into Palantir's Maven system - which speeds up military decision-making. The military's use of mass pattern recognition, Hall said, was already helping cut the long hours of human intelligence resources. But the development that would significantly impact the digital battlefield would be the concept of RSI, recursive self-improvement. Currently, most AI models rely on large language models (LLMs) that generate text on externally trained datasets. The development of an RSI model would do it on its own, without human intervention. 'It's a theoretical construct and it's not been hit yet, but it's this idea that there could be and will be a point at which AI is able to update and improve itself recursively. That is to say, autonomously,' Hall said.  U.S. Secretary of War Pete Hegseth provides updates on military operations in Iran, where AI systems are currently being used to sift through target data and surveillance  Currently most AI systems use large language models (LLMs) trained by humans to generate text  RSI modeled AI could be integrated into military robots, drones, and other killing and hacking machines.  Making that technological leap, Hall said, would give a country an enormous edge in warfare. 'The country that has that will have full-spectrum battlefield dominance in things like encryption, in things like code hacking, hacking into missile systems, hacking into infrastructure, and also in cybersecurity,' Hall said. If China beat the United States to RSI, they would likely us it to develop a totalitarian digital world, Hall warned, something that both Democrats and Republicans should work together to prevent.  'None of us want to live in a CCP surveillance state or a techno-authoritarian regime, I don't care what a person's background is politically, I don't think any American wants that and recognizes that that's a real problem,' he said. The use of AI in the military in the United States has relied on the development of the technology by private industry in the Silicon Valley tech industry. But recently the use of AI in the military has hit a snag, after Secretary of War Pete Hegseth demanded full control of Anthropic's AI technology systems for lawful military use. The fight was triggered by Anthropic CEO Dario Amodei who set red lines for the use of their technology to ban the Pentagon from using it for fully autonomous lethal systems or mass surveillance of US citizens. Hegseth berated Amodei for what he considered a 'master class in arrogance and betrayal,' and said the Pentagon would not work with Anthropic's 'woke' restrictions. Anthropic CEO Dario Amodei has tried to set limits for the way the Pentagon uses their technology  President Donald Trump and Pete Hegseth challenged Anthropic CEO Dario Amodei for trying to limit the Pentagon's use of their AI technology Hegseth confirmed the Department of War only contract with AI companies who accede to 'any lawful use' of the technology, and threatened to label Anthropic as a supply chain risk, if they did not agree to their terms. Trump directed the United States to root out Anthropic's technology from the Pentagon, describing them as a 'radical left, woke company.' Hall said that the Department of War correctly stood up for their independence over a private company, despite Anthropic's attempts to place limitations on their technology. 'The terms of service agreement from a vendor government contract should not override the terms of the battlefield that a commander-in-chief should, no matter who they are, whether it's a Democrat president or a Republican president' Hall said. Constitutionally, Hall noted, the President and the military were elected by the American to make the decisions about war and with American lives on the line, no company should be in a position where they influence or control military operations. 'I think that any commander-in-chief has to be able to know that in the middle of a special operations, you know, secret, you know, mission, that a guardrail from a technologies term of service agreement isn't going to throw up a red gate and, you know, disconnect when we have human lives on the line,' he said. Any military contract with a private company includes the direction that the products will be used for all lawful purposes by the United States. He warned that long-standing efforts by some companies in the tech community to restrict their work for the Pentagon would harm the overall race with China to compete with AI. 'The soldiers and sailors and airmen and marines that give you the physical security that made you billionaires and endlessly wealthy, they were there for you. It's time for you to be there for them,' he said. No comments have so far been submitted. Why not be the first to send us your thoughts, or debate this issue live on our message boards. By posting your comment you agree to our house rules. Do you want to automatically post your MailOnline comments to your Facebook Timeline? Your comment will be posted to MailOnline as usual. Do you want to automatically post your MailOnline comments to your Facebook Timeline? Your comment will be posted to MailOnline as usual We will automatically post your comment and a link to the news story to your Facebook timeline at the same time it is posted on MailOnline. To do this we will link your MailOnline account with your Facebook account. We’ll ask you to confirm this for your first post to Facebook. You can choose on each post whether you would like it to be posted to Facebook. Your details from Facebook will be used to provide you with tailored content, marketing and ads in line with our Privacy Policy.