Post #2951548

@samwilson@wikis.world this! we are in the process of evaluating ML transcription in scanned documents and feature identification in photos and I've been arguing that no, we really need to be absolutely clear on where the information from any particular field was sourced and whether it was later reviewed ...