Mohammad Hosseini: Why does the success of Chinese AI invention DeepSeek elicit fear?

Chicago Tribune

05-02-2025

The artificial intelligence boom led by American tech bros needed one instance of serious competition from China, namely DeepSeek, to shiver in fear and lose $94 billion of wealth in a single day. This cannot be because the AI industry is afraid of more players entering the game. There is already healthy competition among prominent AI companies such as Amazon, Google, Meta and OpenAI — all of which are U.S.-based.
Could it be because of DeepSeek's performance? Probably not, because initial experiments show that DeepSeek's performance for most users is on a par with ChatGPT. But again, Gemini and Claude are also on a par with ChatGPT so this is not new either. If DeepSeek offers good value at a reasonable price and was developed using significantly fewer chips and is more efficient to run, resulting in a reduced environmental impact, shouldn't we all be happy?
Perhaps what makes DeepSeek different is that it came from China.
If we believe in free markets and the survival of the most efficient processes, it is only fair to consider DeepSeek a success story in technology, business and long-term vision. However, because of DeepSeek's country of origin, some fail to see it as a success story. This is a clear violation of modern sensibilities regarding respect for others and their achievements and demonstrates traces of conspiracy theories that will ultimately backfire and hinder the success of the American technology sector. As long as China is seen as an enemy on all fronts, and not a successful competitor in the technology sector with long-term vision and plans, the American tech sector will fail to recognize its own flaws and shortcomings.
China has shaped its foreign policy strategy toward the U.S. primarily around technological competition. Unlike Russia, which prioritizes military capabilities, or the European Union, which focuses on economic rivalry, China sees technology as its key battleground. And to their credit, when exploring the facts rather than rhetoric, it is evident that China is executing this strategy efficiently and deliberately. For example, China's investment in technology infrastructure has increased significantly since 2013, when the Made in China 2025, or MIC25, initiative was adopted. MIC25 is a 10-year ambitious plan aimed at catapulting China into a global leader in high-tech industries by reducing reliance on foreign technology and developing 10 high-tech industries such as AI, advanced robotics, telecommunication and aerospace engineering.
As such, DeepSeek's success should be seen as part of a much bigger trend that will challenge the U.S. and its technology sector in the next few years. Reflecting only the perspective of technology development, whether the Chinese developers used sanctioned Nvidia chips or employed ChatGPT's design and content to train DeepSeek are both secondary. Especially given that OpenAI, Amazon, Meta, X and other American AI developers do not have clean legal, social and environmental records either. Furthermore, ethical standards of American AI companies — especially in terms of used sources, transparency and reliability — have been questioned for a long time, but until now not much has changed, as demonstrated by the ongoing New York Times lawsuit against OpenAI.
From a business and financial point of view, losing clientele to DeepSeek should be investigated systematically with a view to understanding user needs, and scapegoating will never help this process. Indeed, antagonizing the competition prevents U.S. companies from learning from their mistakes. The point that matters the most about DeepSeek's success is that it is much cheaper, uses less energy and offers open model architecture and open weights — both of which are treated as trade secrets in systems like ChatGPT and Gemini.
To better understand DeepSeek and China's success from a technology perspective, let's consider a technology that was championed in the U.S., namely the internet. In the 1970s and 1980s, networking technologies were being developed globally, but many were proprietary and costly, such as France's Minitel system, which provided services such as online banking and shopping before the internet. However, the U.S. Department of Defense (through DARPA — Defense Advanced Research Projects Agency) pioneered Transmission Control Protocol/Internet Protocol, known as TCP/IP, which became the foundation of the modern internet. Unlike the proprietary systems, TCP/IP was openly shared and freely modifiable, enabling rapid adoption, innovation and decentralization. This openness led to the explosive growth of the internet and allowed U.S. companies — such as Cisco, Microsoft and later Amazon and Google — to dominate the digital economy.
One might interpret the success of TCP/IP as part of a conspiracy, a manifestation of American hegemony and military ambitions, suggesting it was part of a long-term plan to dominate and manipulate global communications. However, from a history of technology perspective, TCP/IP represents a remarkable success story, which was developed by visionaries. Indeed, developers of TCP/IP had a clear vision and worked hard to turn it into an international success. DeepSeek, too, is a technological success story and should be understood in this light. Only then can the American technology sector learn from its mistakes and the competition, and succeed.
Mohammad Hosseini, Ph.D., is an assistant professor in the Department of Preventive Medicine at Northwestern University's Feinberg School of Medicine.