Survey: AI security incidents rise as companies deepen integration

A new survey by device management company Jamf finds that companies using artificial intelligence extensively are significantly more likely to experience security incidents tied to that use. The survey polled 687 IT professionals and highlights a growing gap between how fast organizations adopt AI and how well they can govern it.

Nearly three quarters of respondents (72.9 percent) said their organization already uses AI in some form. Companies with deep AI integration face a 40 percent higher likelihood of experiencing an AI-related security incident compared to those in early adoption stages. The report attributes this to a lack of oversight: governance capabilities tend to lag behind adoption.

Twenty-two percent of respondents said their organization had already experienced an AI-related security incident involving unexpected costs, a security problem, or both. Another 59.7 percent described such incidents as an immediate business risk.

Four challenges organizations face

• Shadow AI: Employees use AI tools without IT approval, leaving organizations with little visibility into how company data is being used.
• Developer and agentic AI: These tools use command-line interfaces, embedded models, and autonomous workflows that traditional monitoring tools often miss.
• Expanding AI features: Software vendors are continuously adding AI capabilities to existing products, increasing the number of tools IT teams must evaluate.
• Cost management: Usage-based pricing and overlapping subscriptions make it difficult to track AI spending and assess actual value.

When asked about priorities for the next twelve months, respondents ranked automating IT processes first (44.4 percent), followed by AI-powered productivity tools (41.0 percent) and establishing AI governance (36.7 percent). The data suggests IT teams are not looking to slow AI adoption but want to enable it responsibly.

Beth Tschida, CEO of Jamf, said: “AI is not a single application that IT can approve, deploy, and set aside. The challenge is maintaining visibility and control over the AI that is increasingly being used.” She added that Apple’s privacy model and built-in management features give IT teams a strong foundation for AI governance.

The report concludes that the central question for organizations is no longer whether to adopt AI, but how to ensure governance keeps pace with that adoption.