U.S. Raises Alarm Over China’s AI Use in Cyber Threats, Sparking Geopolitical Tensions
WASHINGTON, D.C. — As the U.S.-China rivalry intensifies, American officials are sounding the alarm over China’s use of artificial intelligence (AI) in activities that could undermine U.S. national security, from cyber espionage to disinformation campaigns. While the U.S. State Department has not explicitly accused China of using AI to “attack” the United States, recent actions by lawmakers and intelligence agencies highlight growing concerns about Beijing’s AI advancements, fueling a narrative of technological warfare that could impact global economies, including tourism hubs like Orlando, Florida.
In March 2025, Representatives Josh Gottheimer and Darin LaHood introduced a bipartisan bill to ban the Chinese AI model DeepSeek on U.S. government devices, citing risks of data collection and surveillance. The lawmakers warned that DeepSeek, developed by a Shanghai-based startup, could feed sensitive information to Chinese authorities, potentially enabling phishing, disinformation, and social media monitoring targeting the U.S. OpenAI’s February 2025 report corroborated these fears, noting that China-based actors used DeepSeek and ChatGPT to generate malicious content, including anti-U.S. articles published in Latin American media.[](https://www.nytimes.com/2025/04/16/technology/nvidia-deepseek-china-ai-trump.html)
The broader context includes U.S. accusations of Chinese state-sponsored cyberattacks. In March 2024, the Justice Department charged seven Chinese hackers linked to the APT31 group for a 14-year espionage campaign targeting U.S. officials and businesses, while a January 2025 Treasury Department breach was attributed to a China-based actor. Though AI was not explicitly cited, FBI Director Christopher Wray has warned since July 2023 that China could leverage stolen U.S. AI technology to enhance cyberattacks, amplifying their scale and precision.[](https://www.scmp.com/tech/big-tech/article/3305663/deepseek-alibaba-help-china-narrow-gap-us-leading-ai-models-stanford-report)
China, meanwhile, has countered with its own accusations, claiming on April 15, 2025, that the U.S. National Security Agency launched “advanced” cyberattacks during the Asian Winter Games, targeting critical infrastructure. Beijing’s state media also released AI-generated content, including a sci-fi video titled “T.A.R.I.F.F.,” rebuking U.S. tariffs and portraying America as destabilizing global trade—a rhetorical escalation in the AI narrative.[](https://www.aljazeera.com/news/2025/4/15/china-accuses-us-of-launching-cyberattacks-during-asian-winter-games)[](https://time.com/7274777/china-trump-tariffs-ai-music-video-short-film-state-media-artificial-intelligence/)
Critically, the term “attack” may overstate China’s AI activities, which primarily involve espionage and influence rather than direct assaults on infrastructure. Both nations have incentives to amplify threats: the U.S. to justify export controls and sanctions, and China to deflect criticism and rally domestic support. Evidence of AI-specific attacks remains scarce, with many claims relying on classified data, raising questions about transparency and potential fearmongering.
In Orlando, home to Universal’s Epic Universe, set to generate $2 billion for Florida in 2025, these tensions could disrupt Chinese tourism and investment, critical to the region’s economy. A deepening trade war, exemplified by Trump’s 145% tariffs on Chinese goods, risks reducing visitor flows, impacting jobs and local businesses.[](https://www.washingtonpost.com/technology/2025/04/12/tariffs-trump-threaten-silicon-valley-ai/)
As the AI race tightens—Stanford’s 2025 AI Index notes China’s models nearing U.S. parity—the U.S. must balance vigilance with diplomacy to avoid escalating tensions that could harm global stability and local economies like Orlando’s. For now, the “AI attack” narrative remains a potent symbol of a broader technological cold war.[](https://www.scmp.com/tech/big-tech/article/3305663/deepseek-alibaba-help-china-narrow-gap-us-leading-ai-models-stanford-report)
—