US AI Anti-China Act: Using DeepSeek is a felony, with a maximum sentence of 20 years and a fine of 1 million pounds.

The US Congress recently proposed the 'Decoupling America’s Artificial Intelligence Capabilities from China Act 2025', which explicitly prohibits Americans from downloading or using the technology of the Chinese AI company DeepSeek. Violators may face up to 20 years in prison and a fine of $1 million. The bill also expands restrictions on AI technology cooperation, with violators potentially facing a fine of $100 million. Non-US citizens involved in technology transfer may also be deported. The US is expanding AI restrictions to prevent Chinese technological development. The bill, proposed by Republican Senator Josh Hawley, who has long taken a tough stance on Chinese tech development and has been involved in investigations into Google and Facebook's alleged election interference, states that individuals or companies in the US cooperating with Chinese universities, research institutions, or laboratories in the development of AI technology may be in violation of the law, with violators facing fines of up to $100 million. The US government had previously banned the use of DeepSeek by government agencies, including the Department of Defense and Congress, and Texas Governor Greg Abbott went even further by signing an executive order banning the use of DeepSeek and the Chinese social media platform Xiaohongshu by state government electronic devices, making it the first in the US to do so. Senator Josh Hawley emphasized that the US must guard against the rapid rise of Chinese AI technology in order to maintain national security and technological leadership. DeepSeek's rise has alerted the US, and the Taiwanese government is following suit in addressing cybersecurity risks. DeepSeek has recently gained attention for its low-cost, high-performance AI models, which have not only impacted Silicon Valley and the tech industry in the US but also raised cybersecurity concerns. The US Navy has already requested all personnel to avoid using DeepSeek-related technology due to 'potential security and ethical issues'. In Taiwan, Premier Su Tseng-chang discussed the cybersecurity risks posed by DeepSeek during a meeting on the 3rd and decided to completely ban the use of the AI service by government agencies to ensure national information security. Experts believe that it may be difficult to fully ban downloads, and Taiwan should strengthen its AI competitiveness. Professor Hong Shih-han of the Department of Computer Science and Information Engineering at National Taiwan University pointed out that because the technology is open source, users can download and run it locally, making it difficult to completely ban. KMT tech legislator Ge Ju-jun believes that Taiwan does not need to speculate or attack DeepSeek, but should learn from its development model and find its own advantages through creativity and technological breakthroughs in the global AI competition.