Rising security budgets leave gaps in AI oversight, DNSFilter survey finds
Many organizations are increasing cybersecurity spending, but stronger budgets have not resolved uncertainty about how artificial intelligence tools are being connected to corporate systems. A report published by DNSFilter on Nov. 4, 2026, found that IT and cybersecurity workers continue to see weaknesses in AI oversight and in the management of security tools.
The findings come from a survey of 400 IT and cybersecurity workers. Three-quarters said their organizations had increased IT and security budgets during the previous 12 months, while roughly one-fifth reported an increase of more than 10%. Respondents also identified AI security and governance as a leading investment priority: four in 10 said it would receive their next security dollar. Cloud security was the second-most-cited priority, at 25%.
AI use is outpacing oversight
Despite that attention, survey responses point to limited visibility into unapproved AI applications. Nearly half of respondents, 46%, said they were confident that unapproved AI apps were connected to their organizations’ systems. One-fifth said an unapproved AI integration had resulted in a cybersecurity incident.
The survey also highlighted gaps in how employees report connections between AI services and work accounts. Thirty-seven percent of respondents said their organizations typically learn about those connections only after they happen. Another 7% said employees are not required to notify IT when connecting an AI app. DNSFilter characterized these situations as decisions about access to corporate data being made without the involvement of the people responsible for managing the risk.
These results do not establish how widespread any particular AI-related incident is across the broader business population. They do, however, describe a practical governance challenge for security teams: a policy or investment priority cannot provide effective oversight if teams do not know which applications are in use, what they connect to, or whether their use has been approved.
Tool management remains uncertain
Respondents reported uncertainty beyond AI. Sixty-two percent said they were unsure whether all of their security tools were properly configured. The same share could not confirm that every tool was actively monitored. That leaves organizations with questions about the operational state of technologies they may already have funded.
The report’s findings distinguish spending from governance. Buying a security product or adopting an AI policy does not, on its own, demonstrate that controls are consistently configured, monitored and applied. For enterprise leaders, the figures suggest that investment decisions need to be connected to clear ownership and visibility across the tools and services operating in the environment.
What security leaders can take from the findings
Organizations can use governance reporting to make the gap between policy and practice easier to assess. Useful measures may include whether AI applications have been reviewed, whether employees know how to request approval, and whether security teams can identify connections to corporate accounts. These are operational considerations prompted by the survey findings, not additional results reported by DNSFilter.
Security and IT teams can also review how they track the configuration and monitoring status of their defensive tools. The survey does not identify particular products or prescribe a specific dashboard, but the reported uncertainty underlines why leaders need a reliable view of tool ownership and operating status. Such visibility can help decision-makers identify where a policy lacks an implementation process or where a purchased control is not receiving adequate attention.
DNSFilter said organizations are adopting AI faster than their governance is developing. The survey suggests that increased spending is not yet translating into confidence that AI use and security controls are under consistent oversight. For CISOs and enterprise IT leaders, the central task is to pair new investment with defined approval processes, accountable owners and ongoing monitoring—so that decisions about connecting AI services to corporate systems do not happen outside established risk management.
