There is nothing intelligent in AI except the human intelligence (HI) that built it. This is particularly true in mineral exploration, where geologists play a vital role in discovery. This talk highlights how transferring qualitative knowledge from geologists into AI systems has led to the discovery of world class copper deposits. Through collaborations with industry leaders, research at Stanford Mineral-X has transformed geological hypotheses from multiple exploration experts into optimized drilling strategies using AI for both greenfield and brownfield exploration. Large ensembles of geological models are generated to ensure geological uncertainty is fully explored in drilling decisions, while maintaining geological realism when integrating Multiphysics datasets. This HI to AI approach enables systematic testing and falsification of existing geological hypotheses that drive mineral exploration. The result is the detection of one of the world largest copper deposit with significantly fewer drillholes and faster discovery timelines. More importantly, improved modeling of subsurface mineralogical heterogeneity also enabled faster planning for subsequent mining and mineral processing. Integrating human intelligence from geologists and geophysicists with AI offers a new paradigm for future exploration decision making.