[Prof_Sai]

You weren't imagining that the AI could output a whole comic on its own, were you? They can't even figure out how the alphabet works!

I was figuring that each panel would need to be assembled from multiple renderings, require extensive touch up, and need to be re-rendered many times to get the right pose and activity. Few AI artists seem to know how to get an AI to consistently create the same people in different poses, but wearing the same clothes. Compositions need to be manually extended and modified to include space for word balloons. Before any of that, we'd need page layouts so we know what to ask the AI for. That could not be automated at all. And of course a concise and efficiently told story that fits in the available number of pages - That's the opposite of what AI stories are like. I haven't seen any sign that anyone has tried to teach an AI about panel transitions or visual storytelling. All of that has to come from the artists & writers.

AI might be able to improve art quality within individual panels, but comics include a lot of hard work beyond that.