The National Institute of Standards and Technology is seeking industry input on practices and methodologies for assessing the secure development and deployment of artificial intelligence agents.
In a request for information posted on the Federal Register Wednesday, NIST said responses will support the mission of the Center for AI Standards and Innovation, or CAISI, to evaluate and develop guidelines to mitigate security risks associated with AI.
The agency specifically wants comments on security threats, technical controls, assessment methods, deployment safeguards and research priorities.
Responses are due March 9.
Gain a deeper understanding of AI implementation and capabilities at the Potomac Officers Club’s 2026 Artificial Intelligence Summit on March 19. Experts from across industry and government will be present at the event to share insights into the rapidly evolving technology. Click here to join.
What Security Risks Are Associated With Agentic AI?
NIST explained in the RFI that AI agents can autonomously plan and take actions that may impact real-world systems. The technology can also be deployed with little to no human oversight.
According to the agency, AI agents face security vulnerabilities also present in more traditional software systems. However, they also introduce novel risks that, if exploited, may pose a threat to public safety and slow down the adoption of AI.
The documents identify hijacking, backdoor attacks and data poisoning as some of the risks associated with agentic AI.
What Is CAISI?
CAISI, housed within NIST, was established to serve as industry’s primary point of contact within the government for evaluating and securing commercial AI, especially capabilities that may pose risks to national security.
In December, the organization posted a call for AI experts to work on initiatives that promote AI safety and competition.
