A University of Cambridge philosopher argues that our evidence for what constitutes consciousness is far too limited to tell if or when artificial intelligence has made the leap—and a valid test for doing so will remain out of reach for the foreseeable future.