That seems easy for AI to adapt to, and has a massive side effect of calling the author's reliability into question. Are those people going to go back and fix the intentional typos in a couple of years once AI spam is also full of typos?