jscalzi@threads.net ("John Scalzi") wrote:
"AI" as it is implemented today will *always* generate errors and "hallucinate," it's baked straight into the model. Pushing out AI-generated podcasts without (apparently) prior editorial oversight is a very fine way for the Post to dilute the credibility it has as a fact-based entity - which the staff very well understands, even if the management doesn't. "AI" sucks for accuracy.