Guardrails for the Bots: Securing AI Agents in Higher Education

AI agents are rapidly moving from experimental pilots to operational tools across higher education, performing tasks such as scheduling, data retrieval, system integration, and decision support. As these agents gain autonomy and access to institutional systems, they introduce new security, privacy, and governance risks that extend beyond traditional application or user-based controls. This session explores how colleges and universities can responsibly secure AI agents while enabling innovation. Participants will examine the unique risk profile of agentic AI, including identity and access management, data exposure, auditability, and human oversight. The discussion will connect emerging AI agent use cases with established higher education security principles such as zero trust, least privilege, privacy by design, and shared governance. Designed for security leaders, IT architects, privacy professionals, and institutional decision-makers, this session provides a practical framework for evaluating, deploying, and governing AI agents in ways that align with institutional values, regulatory obligations, and risk tolerance. Attendees will leave with actionable guidance to move beyond ad hoc experimentation toward a sustainable and secure approach to agentic AI in higher education.