The main pattern is a strong trust gradient, as you can see in the figure below. People who trust news more are much more likely to think AI-made news would be more accurate, unbiased, informative, easier to understand, and up to date, while people who strongly distrust the news are notably more negative on those same qualities. The clearest exception is that AI will make news “Cheaper to make”, where everyone seems to assume that this will be exactly what happens and news’ general trustworthiness, where even those trusting news the most think that AI will make news less trustworthy.