China’s DeepSeek artificial intelligence shakes up markets and raises national security concerns

Wall Street started the week in a cold sweat, thanks to DeepSeek, a little-known Chinese artificial intelligence lab that just dropped a bombshell: a lightning-fast, affordable large-scale language model that completely Shocked Silicon Valley. The Hangzhou-based company claims it developed the product in just two months, cost less than $6 million, and uses low-performance chips from Nvidia (NVDA). The company’s shares fell more than 15% early on Monday (January 27). If the new company, launched in mid-2023, can produce reliable artificial intelligence models that are cheaper and faster than existing applications, it would be a stunning blow to U.S. companies such as OpenAI, Anthropic and Meta (META) . DeepSeek’s new artificial intelligence application is already available online, including in Apple’s App Store, and is moving quickly.

The timing couldn’t be worse for American businesses, given President Donald Trump’s bold actions. announcement Last week, a new initiative called Stargate AI was launched involving OpenAI, SoftBank (SFTBF) and Oracle, which Trump promised would secure America’s “technological future.” Creating hundreds of thousands of jobs in the process. Has that hopeful vision faded?

It’s too early to tell, but it’s now clear that the United States and China are in a fierce competition to harness the power of artificial intelligence for the world. One cannot ignore the comparisons to the Cold War “space race” between the United States and the Soviet Union, with many comparing DeepSeek’s innovation to the Soviet Union’s launch of its first artificial satellite, Sputnik, in 1957, which shocked Americans as they realized that Moscow had entered before us. space.

Like Sputnik, DeepSeek’s claimed advances have alarming national security implications. While Wall Street worries about valuations, the Pentagon is worried about China’s advances in artificial intelligence, which could give the People’s Liberation Army a war advantage. Given how senior U.S. military leaders Evaluate It’s not an idle threat that our increasingly public conflict with Beijing could turn into a pitched battle during Trump’s second term.

The Pentagon’s interest in artificial intelligence and its military implications is not new. Five years ago, the Department of Defense Joint Artificial Intelligence Center expand Support operational plans, not just test new technologies. Last year, Craig Martell, The top Pentagon artificial intelligence official says the new technology is revolutionizing the way the Defense Department does business. as martell explained: “Imagine a world where combatant commanders can see everything they need to make strategic decisions…imagine a world where combatant commanders don’t have access to this information via PowerPoint or email from around the world . [organization] –The turnaround time for situational awareness has been reduced from one or two days to ten minutes.

While the much-vaunted “fog of war” will never completely dissipate, artificial intelligence promises to clear it faster and with a speed and precision unprecedented in the history of warfare. Artificial intelligence will quickly and accurately reduce the information burden on military personnel, providing a tighter “decision-making loop” for U.S. generals and admirals. If DeepSeek’s innovation lives up to its hype, Beijing may have gained a decisive advantage that would allow the People’s Liberation Army to outthink and outmaneuver the U.S. military in any confrontation in the Western Pacific, most likely over Taiwan.

The U.S. intelligence community is also concerned about China’s progress in artificial intelligence. U.S. spy agencies have spent significant amounts of money in recent years to determine how to use artificial intelligence to improve the speed and accuracy of intelligence assessments. Artificial intelligence can suppress the “information fire hose” that hinders rapid analysis of complex intelligence problems, using technology to make human assessments faster and more accurate.

last summer, Lakshmi Raman, CIA’s top official on artificial intelligence, illuminated How does this work. “Imagine all the news reports that come in from around the world every minute of every day. That’s also the data that we have, and we have to use artificial intelligence to help us sort through it,” she explained, adding that artificial intelligence is facilitating the CIA’s secrets Intelligence gathering, not just analysis.

The NSA has also embraced AI with enthusiasm, recognizing its game-changing potential in sifting through vast amounts of collected intelligence data to find puzzle pieces of national security value. The NSA is also protecting the United States from foreign artificial intelligence programs, and the agency recently established Artificial Intelligence Security Center. As NSA Director Timothy Haugh said: “When an enterprise runs AI systems, it faces new attack surfaces in the AI ​​development lifecycle and AI capabilities in model inference services. We need to protect these systems now Protect yourself from threats and reduce risks.

DeepSeek’s claimed advances further exacerbate threats from foreign artificial intelligence. The good news is that no one is sure how real China’s AI advances are. Communists often lie. The Soviet success of Sputnik proved illusory when Moscow launched Yuri Gagarin into space in 1961 (a month before the United States). The Soviet space program was hampered by quality and safety issues, and despite early Kremlin propaganda, the United States won the space race with the 1969 moon landing. The bad news is that time has passed.

John R. Schindler serves as a senior intelligence analyst and counterintelligence officer at the National Security Agency