AI and Cybersecurity in the Public Sector for the Non-Expert

AI is creating new security challenges for the government. Public employees may inadvertently expose sensitive information through AI tools. Vendors may use government data in unexpected ways. Public-facing AI systems can create new vulnerabilities, operational risks, and public trust concerns.

Increasingly, these decisions are not being made by cybersecurity teams alone. Program managers, analysts, procurement officials, communications staff, and agency leaders are making choices about AI tools, vendors, workflows, and public-facing services that can have significant security implications.

This series focuses on the practical cybersecurity questions faced by non-technical public-sector professionals. Participants will learn how AI changes risk, what questions to ask when adopting AI tools and services, and how to work effectively with cybersecurity, privacy, legal, procurement, and technology teams. 

The series provides practical guidance for identifying risks, reducing vulnerabilities, and supporting responsible AI adoption.

Designed for state and local government program managers, analysts, project leads, and agency staff, the series assumes no technical or cybersecurity background.

By the end of this series, participants will be able to:

• Explain how AI-related security risks can affect public services, operations, and public trust.
• Identify common AI-related security and operational risks, including sensitive data exposure, shadow AI, vendor access, prompt injection, inaccurate outputs, and staff overreliance.
• Recognize when to engage IT, cybersecurity, privacy, legal, procurement, and communications colleagues in AI-related decisions.
• Ask practical questions about data use, public-facing services, staff workflows, vendor tools, and escalation paths.
• Distinguish between cybersecurity responsibilities owned by technical teams and risk-reduction actions that program managers can support through program design, workflows, and staff practices.