Calls for AI to be classified as critical infrastructure amid growing security concerns

A new report from the nonprofit Americans for Responsible Innovation has urged the U.S. federal government to classify key aspects of the artificial intelligence sector as critical infrastructure. The report argues that AI systems are now deeply integrated into public and private services, making them a vital component of national security and economic stability. It calls for the Cybersecurity and Infrastructure Security Agency (CISA) to lead cybersecurity efforts for the sector, citing its experience managing other critical infrastructure areas and its statutory mission to address cross-sector cyber threats. The report defines the AI sector broadly, including frontier models, data centers, specialized hardware, and the platforms used to deploy AI at scale. Authors Terrence Kelly and Jessica Maksimov highlight that an attack on AI infrastructure could have cascading effects across multiple sectors, emphasizing the need for coordinated federal oversight. Maksimov noted that CISA’s ability to coordinate across different departments and industries makes it a strong candidate for leading AI cybersecurity efforts. The U.S. is particularly vulnerable to AI supply chain disruptions due to the concentration of frontier AI companies and computing resources within the country. Recent attacks, such as Iranian drones targeting Amazon data centers and Ukrainian drones striking Russian e-commerce platforms, have demonstrated the potential for significant disruptions to critical internet services. Maksimov warned that U.S. AI infrastructure is highly valuable to attackers and thus remains a prime target for both physical and cyber threats. Currently, 16 critical infrastructure sectors are managed by the federal government, and designation carries significant weight in terms of resource allocation and prioritization. Matt Hayden, a former assistant secretary of homeland security, explained that being classified as critical infrastructure places an entity in a special category, granting access to federal tools and resources such as cybersecurity software, threat intelligence, and incident response services. He also noted that while the AI sector may eventually be designated, other sectors like space and cloud computing have similarly sought such recognition. Bob Kolasky, former director of the National Risk Management Center at CISA, believes companies like OpenAI and Anthropic, as well as data center operators, will eventually be classified as critical infrastructure. However, he acknowledged that the process of designating new sectors is complex and often leads to bureaucratic challenges. Kolasky also pointed out that the AI sector may bring unique coordination issues, as the existing 16 sectors function in varied ways. New programs like ANCHOR-CI, introduced by the Department of Homeland Security, aim to streamline the process by allowing CISA to convene ad-hoc meetings and add companies to existing sectors. Despite the potential benefits of designating AI as critical infrastructure, experts warn that the process could result in significant bureaucratic conflicts, as seen in past attempts to classify space-based infrastructure. The Trump administration’s influence on AI policy remains a factor, with the Departments of Commerce and Treasury playing prominent roles in shaping regulations. As the AI industry continues to evolve rapidly, the push for federal oversight reflects growing concerns about its role in national security and economic resilience.