Ironically if you explain to those talk show hosts how they are wrong about how ChatGPT learns (or doesn't learn) and use all the right arguments and proofs so that they finally concede, chances are that they too won't quite learn from that and keep repeating their previous bias next time.