More than 20% of organizations running macOS networks have lost money or suffered a cyberattack tied to their use of AI tools, according to Jamf’s survey of 687 IT and security leaders. Around 73% of those enterprises have already deployed AI, and another 20% plan to do so. Yet nearly six in ten expect an AI-related incident soon. The most common threat is “shadow AI,” where employees adopt unapproved tools that IT teams can’t track. Without visibility, security and governance break down fast. Admins also worry about agentic AI—automated systems that make decisions on their own. They say it’s hard to give these agents the right permissions without risking sensitive code or data. Then there’s the vendor jumble: each new AI tool needs vetting, and the pace of innovation leaves IT teams racing to keep up. On top of that, usage-based pricing can trigger surprise bills when no one’s monitoring which licenses they’ve bought. Despite these risks, governance ranked only third on respondents’ priority list, behind automating IT tasks and boosting worker productivity, with security improvements in fifth place. Jamf recommends regular audits, strict data-access policies, early-stage governance integration and using built-in tools to reduce complexity.