What UK IT Managers Are Actually Doing About Shadow AI
Shadow AI is employees using unapproved AI tools with work data, and the public advice on it is mostly written by vendors. A thread of 226 replies from IT managers is a better record. It produces three findings worth acting on: organisations that block and then supply a safe alternative frequently find the alternative rejected and shadow use resuming; the real obstacle in most organisations is that nobody will sign a policy because signing means owning the risk; and every control described assumes AI is a destination you can block, which stops being true once search itself is an AI interface. This sets out the arguments on both sides, the tooling practitioners name, the data protection problem created by inspecting prompts, and what to do next.
