A GPT-2 written essay was submitted to the Economist's youth essay contest. One judge, who did not know the essay was written by an AI, gave this review: "It is strongly worded and backs up claims with evidence, but the idea is not incredibly original." economist.com/open-future/20…

That makes sense. This system just learns a mapping to the feature space of human conversations, averaged over all available data. It makes perfect sense that it would just repeat the most generic (boring) points for a topic.