Policing AI use by counting ‘telltale’ words is flawed and damaging Making people paranoid about employing familiar and useful words is not the way to encourage responsible AI use, say Lilian Schofield and Xue Zhou By Lilian Schofield 3 May