The US Department of Homeland Security has created an AI Safety and Security Board to tackle AI-related security and privacy issues.

This initiative is a response to a 2024 threat assessment that highlighted the risk of AI-assisted cyberattacks on US critical infrastructure.

New security guidelines have been issued to safeguard critical infrastructure against AI-driven threats, drawing from the National Institute of Standards and Technology's AI Risk Management Framework.

AI risks are now divided into three categories: attacks on infrastructure, assaults on AI systems, and failures in AI design.

The Cybersecurity and Infrastructure Security Agency (CISA) has incorporated the AI Risk Management Framework into guidelines for critical infrastructure entities to enhance AI safety and promote risk management culture.