Great approach, but is this viable in the time of ai, like if someone try to fool using ai generated content