The "Advanced AI Security Readiness Act" mandates the Director of the National Security Agency (NSA), through its Artificial Intelligence Security Center, to create an "AI Security Playbook." The primary goal of this playbook is to develop comprehensive strategies for defending "covered AI technologies" from technology theft perpetrated by "threat actors." Covered AI technologies are defined as advanced AI systems, whether developed by the private sector or government, that pose a grave national security threat if acquired or stolen, such as those matching or exceeding human expert performance in critical areas like chemical, biological, radiological, and nuclear matters or cyber offense. The playbook must include several key elements, such as identifying potential vulnerabilities in advanced AI data centers and among developers, focusing on cybersecurity risks unique to AI. It will also pinpoint critical components or information, like AI models and core insights, that would significantly aid threat actors if stolen. Furthermore, the playbook will detail strategies to detect, prevent, and respond to cyber threats targeting these technologies, and analyze the levels of security that might necessitate substantial United States Government involvement in the development or oversight of highly advanced AI systems. In developing the playbook, the NSA Director is required to engage with prominent AI developers and researchers to assess and anticipate the capabilities of advanced AI systems relevant to national security. This engagement includes reviewing industry documents, conducting interviews, and hosting discussions. The Director must also collaborate with a federally funded research and development center experienced in securing AI models from nation-state actors. The final playbook will include both a classified annex and an unclassified portion with general guidelines and best practices for broader dissemination, including to the private sector, with initial and final reports submitted to Congress within 90 and 270 days, respectively.
Get AI-generated questions to help you understand this bill better
Timeline
Introduced in House
Referred to the House Committee on Intelligence (Permanent Select).
Introduced in House
Referred to the House Committee on Intelligence (Permanent Select).
Armed Forces and National Security
Advanced AI Security Readiness Act
USA119th CongressHR-3919| House
| Updated: 6/11/2025
The "Advanced AI Security Readiness Act" mandates the Director of the National Security Agency (NSA), through its Artificial Intelligence Security Center, to create an "AI Security Playbook." The primary goal of this playbook is to develop comprehensive strategies for defending "covered AI technologies" from technology theft perpetrated by "threat actors." Covered AI technologies are defined as advanced AI systems, whether developed by the private sector or government, that pose a grave national security threat if acquired or stolen, such as those matching or exceeding human expert performance in critical areas like chemical, biological, radiological, and nuclear matters or cyber offense. The playbook must include several key elements, such as identifying potential vulnerabilities in advanced AI data centers and among developers, focusing on cybersecurity risks unique to AI. It will also pinpoint critical components or information, like AI models and core insights, that would significantly aid threat actors if stolen. Furthermore, the playbook will detail strategies to detect, prevent, and respond to cyber threats targeting these technologies, and analyze the levels of security that might necessitate substantial United States Government involvement in the development or oversight of highly advanced AI systems. In developing the playbook, the NSA Director is required to engage with prominent AI developers and researchers to assess and anticipate the capabilities of advanced AI systems relevant to national security. This engagement includes reviewing industry documents, conducting interviews, and hosting discussions. The Director must also collaborate with a federally funded research and development center experienced in securing AI models from nation-state actors. The final playbook will include both a classified annex and an unclassified portion with general guidelines and best practices for broader dissemination, including to the private sector, with initial and final reports submitted to Congress within 90 and 270 days, respectively.