# NIST Highlights Federal AI Agent Controls

National Institute of Standards and Technology (NIST) representatives Ryan Galluzzo and Bill Fisher discussed how federal agencies can identify and govern AI agents that access data and act within government systems. The discussion covered agent identity and credentials, zero-trust and least-privilege access, accountability, risks from untrusted data and prompt injection, and preserving human decision authority. It describes a federal AI security and governance concern, not a solicitation, contract award, or specific procurement opportunity.

- Contractors developing or integrating AI for federal environments should assess how their systems support agent-specific identity, appropriately scoped permissions, and auditable actions.
- Agencies and industry stakeholders may need to address prompt-injection and untrusted-data risks, approved-use policies, and human oversight when designing or evaluating AI capabilities.
- The discussion identifies NIST resources but provides no specific contract vehicle, funding, compliance deadline, or solicitation to pursue.

**Jurisdictions:** federal
**Industries:** Information Technology
**Topics:** Artificial Intelligence, Cybersecurity
**Published:** October 15, 2026

### Government Entities
- National Institute of Standards and Technology (NIST)

### Sources
- [
        Where’s your PIV card, HAL? What happens when AI gets an employee ID - Federal News Network](https://federalnewsnetwork.com/podcast/fed-thread-podcast/wheres-your-piv-card-hal-what-happens-when-ai-gets-an-employee-id) - Federal News Network
- [
        Where’s your PIV card, HAL? What happens when AI gets an employee ID - Federal News Network](https://federalnewsnetwork.com/fed-thread/2026/10/wheres-your-piv-card-hal-what-happens-when-ai-gets-an-employee-id) - Federal News Network