Enterprise AI Security: 80% lack IT oversight, risk data loss

Four out of five enterprise AI tools are operating without central IT oversight. This creates significant data security risks for businesses. A new report found that 62% of these AI tools can both read internal files and connect to the internet, opening a path for data to leave the company without authorization. Unlike traditional software, these AI agents embed themselves deeply within company systems, inheriting user permissions and access. This allows them to combine capabilities in unexpected ways, potentially traversing networks in ways IT administrators never intended. Experts warn that autonomous agents with active credentials could modify source code or alter production databases. The recommendation is to focus on visibility and policy controls for these AI tools, rather than trying to block their use entirely. This approach aims to manage risk without slowing down employee workflows. This is crucial because businesses need to secure their data as AI adoption continues to grow.