This! I did play around with ChatGPT for a while and found its responses to be incredibly generic. While they seem impressive at the first glance, in many cases it seemed like the responses seem like the first google result with some keywords replaced.

What’s impressive to me about ChatGPT isn’t the text output but the context awareness of the model.

Not just generic, but often false. It will get even easier to spread false news even if that's not the main goal of the publisher.