So has anyone looked into the issues involved in explaining this level of causation to any AI systems? It seems this might be work, but interesting.