Bridging The AI Governance Gap: Why Deep Integration Is Driving A Surge In Cybersecurity Incidents

As businesses in Australia and New Zealand push deeper into generative and agentic AI, Jamf CEO Beth Tschida says the risk picture has turned counterintuitive: more mature AI deployments can bring more security and cost incidents. Jamf’s global survey of 687 IT and security leaders at Apple-first organizations found 72.9% have deployed AI, while 81.7% either have experienced an AI-related incident or expect one soon. Companies with deeply integrated AI reported incidents 40% more often than those still exploring (27.1% vs. 19.4%). Tschida links the problem to local AI agents running via command-line interfaces and MCP servers on Apple Silicon that traditional network tools can’t observe. Jamf says it has launched “AI Governance” general availability for Mac, using Apple’s Declarative Device Management to detect “Shadow AI” and enforce compliant profiles, and to integrate endpoint agents with Okta for AI Agents.