Sometimes AI makes decisions behind the scenes, where the procedures used are opaque – especially for non-experts. In mission critical cases AI must be open to inspection. This is true in public forums, medical procedures, logistics, and custodial care. Our aim is a definitive format to record decisions (and the procedures and variables that were utilized). This can be accomplished by creating a timestamped block-chain or DLT event. These can be made to create trust so that human error and machine error may be understood. AI reliability must be carefully tested and sand-boxed prior to use. However, after the AI is put into service this solution offers security, operational certainty, and a chain of custody. Sentiment analysis is a straightforward application of Natural Language processing that has been largely standardized. This makes a limited project suitable for this funding. Our black box is like an AI flight recorder rather than a hidden mysterious process.