Tensorization of Fisher information

Fisher information tensorizes when the information in an $n$-observation model is the sum of the information contributions from its observations. For identically distributed observations this meansIndependence and the mean-zero score identity make the cross terms vanish.