This week, members of the U.S. House Homeland Security Committee and the specially established “Committee on CCP Issues” participated in a bipartisan war game simulation to discuss AI-driven cyber threats, resilience of critical infrastructure, and potential threats to U.S. homeland security once the Chinese Communist Party (CCP) gains advanced AI technology.
According to a press release issued on Friday (24th) by the House Homeland Security Committee, the simulation led by the Center for Strategic and International Studies (CSIS) focused on CCP-related hackers launching AI cyberattacks targeting U.S. ports, freight railways, maritime surveillance systems, log systems, and the transportation networks required for deploying U.S. forces if the CCP invades Taiwan.
During the simulation, a specially designed AI agent program simulated coordinated cyber attacks against U.S. assets and critical infrastructure, prompting lawmakers playing the roles of U.S. officials and policy advisors to devise strategic responses.
Former Chairman of the Homeland Security Committee and Republican Congressman Michael McCaul participated in this closed-door war game simulation. He stated that for years, CCP-supported hackers have been seeking vulnerabilities to attack U.S. government and civilian networks, and now advanced AI models exacerbate these cyber risks.
McCaul said, “After hearing multiple briefings from the private sector on AI and its security risks, this war game simulation became an important opportunity for the committee to reassess the nature of warfare, highlighting that cyberspace has become a frontline battlefield on par with traditional military operations.”
According to a briefing provided to reporters before the simulation began, the exercise was set against the backdrop of a large-scale military exercise by the CCP in the summer of 2027 near Taiwan. At that time, U.S. officials were unable to determine whether the Beijing regime intended to exert pressure, implement a de facto blockade, or prepare for a full-scale invasion of Taiwan.
In this scenario, although Beijing avoided using the term “blockade,” its actions effectively isolated Taiwan from the outside world. Subsequently, members of Congress learned that U.S. intelligence analysts suspected CCP hackers had begun targeting ports, freight railway systems, airports, and other transportation networks used to mobilize U.S. forces and supplies. Some attacks caused immediate disruptions, while others seemed to aim at preserving channels for future use.
McCaul discussed key points after the simulation concluded. He emphasized that the focus of this war game simulation was on the CCP, as it is the United States’ “primary adversary in the cyber domain.”
McCaul stated, “The committee has received briefings on the latest AI capabilities in the United States, especially Anthropic’s Mythos model, and the potential for immense destructive power if this model falls into enemy hands.”
The House Homeland Security Committee continues to engage in roundtable meetings, briefings, and hearings to exchange views with the private sector on the future development of cybersecurity and the development of secure AI technologies.
In May of this year, Anthropic presented one of the first live demonstrations of the “Mythos” system to members of the committee, allowing them to witness firsthand how advanced AI detects software vulnerabilities, further highlighting the urgent need for federal agencies to use advanced U.S. models to identify and patch vulnerabilities.
In April, Homeland Security Committee Chairman Andrew R. Garbarino and “Committee on CCP Issues” Chairman John Moolenaar launched a joint investigation into the increasingly widespread adoption of AI models developed by the CCP.
This investigation aims to clarify how the CCP is narrowing the gap in AI innovation with top U.S. labs through model distillation, intellectual property theft, and other illegal means, while also reviewing the potential network security, data security, and critical infrastructure risks posed by the deployment of CCP-developed AI systems in American products and services.