4 days ago
- Business
- South China Morning Post
China and US are locked in high-stakes fight to shape AI future
China and the United States are locked in a fierce battle over artificial intelligence. But the two scientific powerhouses have very different endgames in mind. While Washington wants to dominate the field with a few close allies, Beijing sees its development as a win-win for all. It has called for
global governance , while Chinese industry and research promote
open sources
In its latest move, Huawei Technologies is
offering its CANN platform – Compute Architecture for Neural Networks – as an open-source software toolkit for programmers and researchers. CANN runs on Huawei's Ascend AI processors and is designed to rival Nvidia's proprietary Compute Unified Device Architecture (CUDA), established as the standard platform around the world.
High-level open-source capabilities, including the popular
DeepSeek , aim to bolster China's tech self-sufficiency by providing alternative platforms for developers to build applications for domestic software. But China also sees a bright future for its open-source technology around the world, especially in developing economies.
At the recent
World Artificial Intelligence Conference in Shanghai, global governance and cooperation was its main theme. It opened as the White House released its own
policy framework that makes American AI dominance the main goal, through deregulation, infrastructure investment and expanding AI exports exclusively to allies in a technological race with China. US President Donald Trump has characterised it as a fight the US started, and one it will win.
Premier Li Qiang warned at the conference that because AI required high concentrations of technology, capital and talent, it could end up becoming an 'exclusive game' for a select few rich nations and their companies. Instead, he has called for global governance to better explore and develop the full potential of AI as well as managing risks. Washington's vision is starkly different. Beijing has also offered a 13-point action plan that includes setting up research laboratories, AI education and training, and joint development of AI programming.
