Debate over logging every AI prompt for safety raises privacy concerns

In a video interview at The Next Endeavor 2026, entrepreneur Dave Blundin proposes that AI systems could log every user prompt at the hardware level as a safety measure. He argues this would only work if strict safeguards were in place, such as requiring something like probable cause before anyone could access those logs.

Interviewer Luke pushes back, questioning whether such logging could create a bigger privacy risk than it solves. The conversation weighs the tradeoffs between monitoring AI use for safety purposes and protecting individual privacy rights.