US, Taiwan Finalizing Thoughts on DeepSeek

On Monday, new US AI sanctions applied fines on users that have been using DeepSeek while Taiwan on the other side of the world had a decision taken on measures against the Chinese developed AI, DeepSeek for cybersecurity reasons.

The new law would impose prison sentences up to 20 years and fines up to $1 million on Americans who violate the US AI exports restrictions. US companies might face fines of up to $100 million if they get caught in collaboration with Chinese AI firms, or for AI research in China. It continues to forbid the export of AI related IP and importing and using AI models originating from China.

The US legislation on AI was passed at a time when Silicon Valley is still in shock after the release of DeepSeek’s first large-scale reasoning model, R1-which many analysts claimed matched the leading technology from OpenAI.

A release that highlighted competitiveness of Chinese AI firms against Trump’s US trade which limits China’s access to advanced semiconductor technology.

Taiwan Launches DeepSeek Ban

Taiwan has taken the step to ban the use of DeepSeek AI in all public agencies, as US AI sanctions take place on China, stating that it has been done for cybersecurity reasons and other threats to data security.

On Monday, Premier Jung-tai Cho hosted a council meeting at which a decision was taken against DeepSeek.

National Science and Technology Council (NSTC) Minister Cheng-wen Wu stated that universities and research institutions can use DeepSeek for academic purposes, but this is to be on isolated, non-networked computers without personal data to avoid risks.

“Our priority is to ensure national cybersecurity while allowing limited academic use under strict conditions,” Wu said.

A special advisory was issued to all infrastructure providers and regulatory bodies in the interest of strict cybersecurity protection.

“The risks associated with transmitting user-uploaded data to China remain unclear,” said Executive Yuan spokesperson Michelle Lee on the government’s purpose to protect sensitive national information.

What’s Coming for AI

When US federal AI legislation targets AI infrastructure export limitations to China, this legislation will include a provision to make it illegal for Americans to download an AI model like DeepSeek R1.

With most open-sourced AI models, it is anybody’s guess how this law can be enforced. But if it makes it into law, then worldwide collaboration on the advancement of AI could significantly get affected with the potential of deeply profound effects on investors with interests in the Chinese AI firm.

Both the effectiveness with which Taiwan cracked down on DeepSeek and the US’s legislative push reveal a fragmented AI world-one in which the interests of national security are now driving policy decisions in companies.

A proposed US AI sanctions legislation would give Americans pause before downloading a Chinese AI model, let alone invest in the companies behind them. Far from becoming law, it sets the scene for tougher AI trade policies that could reset the future of the industry.

AI models with no restrictions are heating up the competition, and countries might just have to choose between technological advancements and expanding digital borders while being mindful of national security.

Inside Telecom provides you with an extensive list of content covering all aspects of the tech industry. Keep an eye on our Tech sections to stay informed and up-to-date with our daily articles.