XPost: alt.security.espionage, alt.privacy.anon-server   
   From: anon@invalid.org   
      
   A recent congressional investigation has concluded that the Chinese   
   artificial intelligence platform DeepSeek poses a serious and growing   
   threat to U.S. national security.   
      
   According to a new report from the House Select Committee on the Strategic   
   Competition Between the United States and the Chinese Communist Party,   
   DeepSeek secretly harvests American user data, censors information   
   according to CCP directives, and was likely built using stolen U.S.   
   technology—all while relying on semiconductor chips that should never have   
   reached China in the first place.   
      
   “DeepSeek represents a profound threat to our nation’s security,” the   
   report warns. “Although it presents itself as just another AI chatbot…   
   closer inspection reveals that the app siphons data back to the People’s   
   Republic of China (PRC), creates security vulnerabilities for its users,   
   and relies on a model that covertly censors and manipulates information   
   pursuant to Chinese law.”   
      
   To address this threat, the report recommends an expansion of export   
   controls, stronger enforcement against Chinese AI platforms, and the   
   creation of a federal whistleblower program to report violations. It also   
   calls for heightened coordination among national security agencies to   
   prevent China from achieving a “strategic surprise” in the AI race.   
      
   How DeepSeek Works—and Why It’s Dangerous   
   The committee’s investigation found that DeepSeek transmits extensive data   
   – including chat history, device details, and even user typing behavior –   
   through back-end infrastructure connected to China Mobile, a state-owned   
   telecom firm designated by the U.S. Department of Defense as a Chinese   
   military asset. China Mobile has been banned from operating in the U.S.   
   since 2019 due to fears that “unauthorized access to customer… data could   
   create irreparable damage to U.S. national security.”   
      
   Cybersecurity analysts also discovered that DeepSeek sends user   
   information with “no meaningful security measures,” raising serious   
   concerns that the system was deliberately designed to make Americans’ data   
   easily accessible to Chinese authorities.   
      
   Moreover, the AI model itself is manipulated to serve Beijing’s strategic   
   interests. DeepSeek censored politically sensitive topics, including   
   democracy, Taiwan, and human rights, in 85 percent of test cases. It   
   doesn’t merely avoid controversial topics; it actively rewrites history   
   and reinforces CCP talking points. “Beijing is not just censoring the   
   internet at home. It is embedding its Great Firewall into platforms   
   Americans use every day,” the report notes.   
      
   Built on Stolen Technology   
   DeepSeek’s capabilities didn’t come from scratch. Congressional   
   investigators found that DeepSeek likely used “model distillation” – a   
   technique that copies reasoning capabilities from other AI models – to   
   replicate U.S. systems like OpenAI’s ChatGPT. OpenAI confirmed to Congress   
   that DeepSeek employees circumvented protections, used fraudulent   
   accounts, and extracted model outputs in violation of OpenAI’s terms of   
   service.   
      
   “Through our review, we found that DeepSeek employees circumvented   
   guardrails in OpenAI’s models… to accelerate the development of advanced   
   model reasoning capabilities at a lower cost,” OpenAI told the committee.   
   This kind of intellectual property theft poses serious challenges for U.S.   
   companies trying to maintain a competitive edge in a high-stakes field.   
      
   Smuggled Chips Fuel the Engine   
   Even more disturbing is how DeepSeek built and trained its model.   
   According to the report, DeepSeek uses tens of thousands of high-powered   
   chips made by Nvidia, including A100s, H800s, and H100s – many of which   
   are subject to strict U.S. export controls. These chips are crucial to   
   building large-scale AI models, and selling them to China without a   
   license is prohibited.   
      
   Yet DeepSeek appears to have acquired many of these chips through illegal   
   channels. Investigators discovered a smuggling network operating out of   
   Singapore, where three individuals – one a Chinese national – were charged   
   for illegally exporting Nvidia chips to China. The network was busted   
   shortly after members of Congress raised alarms about chip smuggling via   
   Singapore.   
      
   This incident underscores a broader problem: U.S. companies and   
   intermediaries are still supplying adversaries with the technological   
   tools needed to match and surpass American capabilities.   
      
   A Pattern of Strategic Deception   
   DeepSeek’s founder, Liang Wenfeng, maintains effective control of the   
   company through a complex corporate structure with ties to High-Flyer   
   Quant, a firm that invested $420 million into DeepSeek’s development. The   
   company operates within a state-subsidized Chinese tech corridor built to   
   realize “Xi Jinping Thought” – the ideological core of the CCP.   
      
   Liang’s connections to military-linked researchers and the state-run   
   Zhejiang Lab highlight how closely Chinese tech innovation is tied to   
   national security goals. The app’s integration with entities like Tencent,   
   Baidu, and ByteDance – each with their own histories of surveillance,   
   censorship, or military affiliation – makes DeepSeek more than just a   
   technological threat. It is, in effect, a tool of geopolitical warfare.   
      
   Lessons from the Cold War   
   Experts interviewed for this piece suggest the U.S. needs to reestablish   
   the kind of strict export control framework that helped contain Soviet   
   technological ambitions during the Cold War. Hans-Günter Förstner, a   
   retired professor of international law who enforced Cold War-era trade   
   controls for West Germany, recalled how the U.S.-led Coordination   
   Committee for Multilateral Export Controls (COCOM) successfully denied   
   Moscow access to critical technologies.   
      
   “The Americans… always emphasized that the priority was to deny access to   
   all knowledge and products related to robotics and space technology,”   
   Förstner said. “President Reagan was entirely correct.”   
      
   Dr. Xiàhóu Li Wei, a former senior CCP official who defected to the West,   
   stressed that the current threat is even greater. “The CCP wanted the West   
   to see China as a completely different entity from the Soviet Union,” he   
   said. “It was a deception. The CCP kept winning until recently.”   
      
   Policy Recommendations: What Congress Must Do   
   To counter this growing threat, the report offers several urgent   
   recommendations:   
      
   Expand export controls to include new chip types like Nvidia’s H20 and   
   improve enforcement mechanisms through whistleblower incentives and   
   bilateral crackdowns on smuggling routes like Singapore.   
   Require U.S. firms to track the end users of advanced chips and software.   
   Mandate security and transparency standards for all AI systems trained on   
   U.S.-origin technology.   
   Prohibit the federal government from using Chinese AI platforms like   
   DeepSeek.   
   The report concludes with a stark warning: “The potential for AI strategic   
      
   [continued in next message]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)