AI adoption rate is positively correlated with the frequency of security incidents, highlighting the need for governance
Network management vendor Jamf released a report on Tuesday, based on interviews with 687 IT and security leaders, revealing the correlation between AI adoption and the frequency of security incidents in macOS environments. The report shows that over 20% of organizations have already suffered losses or attacks due to AI tools, and about 60% expect AI-related incidents to occur in the near future. Organizations with deep AI integration have an incident rate as high as 27%, while those in the exploration phase have a rate below 20%. Jamf recommends reducing risks through regular audits, software governance, and early integration of governance frameworks.
Quick Overview
- According to a report released on Tuesday by network management vendor Jamfreport, more than one-fifth of organizations running macOS networks have suffered financial losses or cyberattacks due to the use of AI tools.
- The survey found that about 60% of macOS-based organizations expect an AI-related incident to occur in the near future.
- Based on interviews with 687 IT and security leaders who manage macOS network environments, the report also describes system administrators' AI implementation priorities, the biggest risk areas they face, and risk mitigation recommendations proposed by Jamf.
In-Depth Insights
According to Jamf's report, macOS-based enterprises are widely adopting AI, with about 73% of respondents saying they have already deployed AI and another 20% exploring deployment. However, network administrators overseeing the AI adoption process are facing a range of potential dangers.
The biggest and most well-known risk isshadow AI, which refers to employees using unauthorized and ungoverned AI tools. Jamf points out that when "IT departments have no visibility into which AI systems are being used," "this lack of visibility makes security and governance difficult, if not impossible, to achieve."
Agentic AIis becoming increasingly popular in enterprises, but it alsocomes with risks. IT and security leaders say they struggle to deploy AI agents "in a way that allows users to use them without putting data at risk." The report states: "With improper permission settings, agentic AI can pose a serious threat to codebases—if unsafe or problematic code is added, or necessary code is deleted, the consequences could be dire."
The diversity of AI vendors also poses challenges for technology leaders responsible for procurement. Jamf states in the report: "Reviewing and deploying every possible AI tool is time-consuming and difficult for IT teams, especially given how rapidly AI is evolving."
AI's usage-based billing models can also lead to unexpected expenses for organizations, andgovernance gapsexacerbate this issue—IT leaders often lack clarity on how many paid licenses the organization maintains or which tools offer the best value for money.
Addressing these challenges requiresstrong AI governance, but governance ranks only third on the priority list of survey respondents, behind automating IT management activities and using AI to boost employee productivity. AI security improvements rank fifth.
As organizations adopt AI more broadly, governance will become more urgent. Jamf's data highlights a correlation between the depth of AI adoption and the likelihood of AI-related incidents. Organizations in the exploratory phase of AI have incident rates below 20% (at the low end of the range), while those with AI "deeply integrated" into workflows have incident rates reaching 27% (at the high end of the range).
To reduce AI risk exposure, Jamf recommends that organizations expand visibility through regular audits, focus on software governance rather than user governance (including strictly enforcing data access policies), integrate governance frameworks as early as possible in the deployment phase, and use built-in tools whenever possible for the smoothest experience.