AI Security Becomes Critical Priority

Nudge Security released new AI agent discovery tools that let companies find and control AI agents their employees create. The warning: 80% of organizations are already seeing risks from AI agents with too much access to company data. This matters because employees are quickly deploying AI agents on platforms like Microsoft Copilot Studio without checking security first.

The new tools help teams discover where agents are built, what data they can reach, and who created them—then push back on risky setups. It's the first major defense against "shadow AI" agents spreading across enterprises.

New Research on AI Behavior

Researchers found something surprising: telling an AI model it's an expert actually makes it worse at handling facts. The persona-based prompting helps with safety but backfires on accuracy—useful to know before you trust AI with important information.

Oracle's New Support Center

Oracle opened an AI Customer Excellence Centre in Sydney to help companies handle rapid AI changes. This signals major vendors are building dedicated support just for AI implementations.

What you need to do: Audit your organization's AI agents now before risks pile up.