This document discusses open, connected, and smart cultural heritage. It proposes using crowdsourcing and machine-human computation to support multiple perspectives in interpreting digital cultural content. Disagreement between human annotators is seen as essential for helping machines with semantic interpretation. The CrowdTruth platform and software is presented as a way to harness perspectives through crowdsourced annotation and analytics. Specific use cases are described for exploring audiovisual archives about historical events in the Netherlands.