# Public-Sector Teams Secure AI Agents

Cybersecurity discussions led by Cloudera’s Carolyn Duby and a related analysis of autonomous AI agents highlight gaps in traditional access controls as agents operate across systems at speed and scale. The recommended controls include task-scoped, temporary permissions, continuous authorization and behavioral monitoring, segmentation, rapid containment or emergency stops, human oversight, and auditable data lineage across cloud, on-premises, hybrid, and cross-border environments. The signals identify no solicitation, award, funding, or procurement contact; their procurement relevance is as an indicator of security and governance capabilities public-sector technology teams and contractors may need to evaluate in AI deployments.

- Contractors developing or integrating AI-agent systems can use these controls to inform security architectures and demonstrate governance, auditability, and data-residency capabilities to public-sector customers.
- Procurement teams evaluating AI-enabled solutions can assess whether access is narrowly scoped and temporary, authorization is continuously checked, activity is monitored, and agents can be rapidly contained.
- The signals point to a cross-environment security concern, but do not identify a specific agency requirement, compliance deadline, or available contract opportunity.

**Jurisdictions:** international
**Industries:** Information Technology
**Topics:** Cybersecurity, Artificial Intelligence
**Published:** October 23, 2026

### Vendors
- Cloudera ()

### Sources
- [Carolyn Duby on AI-agent security and zero trust](https://technode.global/2026/10/22/cloudera-carolyn-duby-agent-access-security-qa) - TNGlobal
- [AI Agents Challenge Corporate Cybersecurity and Zero Trust](https://ceo.com.pl/en/ai-agents-corporate-security-zero-trust-access-control-38906) - CEO Magazyn