RNA editing is a posttranscriptional modification that can lead to a change in the encoded protein sequence of a gene. Although a few cases of mammalian coding RNA editing are known to be functionally important, the vast majority of over 2,000 A-to-I editing sites that have been identified from the coding regions of the human genome are likely nonadaptive, representing tolerable promiscuous targeting of editing enzymes. Finding the potentially tiny fraction of beneficial editing sites from the sea of mostly nearly neutral editing is a difficult but important task. Here, we propose and provide evidence that evolutionarily conserved or "hardwired" residues that experience high-level nonsynonymous RNA editing in a species are enriched with beneficial editing. This simple approach allows the prediction of sites where RNA editing is functionally important. We suggest that priority be given to these candidates in future characterizations of the functional and fitness consequences of RNA editing.