On the "generate storyboards" dimension, the images generated for these generated scripts aren't really storyboards, they're visual cues for people who don't read.<p>Amazon launched the first AI "storyboarding" tool about 10 years ago[0]. It read through your script, saw there was a scene where a guy goes to the hospital, and put a picture of a guy in front of a hospital in the storyboards. That was clever, but not useful and served no purpose.<p>The storyboard doesn't exist to pander to people who won't read the script. It's there to help the director, producer, and crew work out how to tell the story. Don't show me a picture of the outside of the hospital because I didn't read the script, show me how the scenes go together so I know whether we need a shot of the outside of the hospital. If we do, where are the actors in frame? Is it a closeup of the actor or a wide establishing shot of the location. In animation, how many backgrounds/locations do we need to tell this story, each of which will need to be designed and drawn or modeled? How many characters are visible in each shot (which translates into how many animators and animator hours will be required).<p>If these images being generated are really actual functioning storyboards intended to show how to translate the story from script to screen for optimum storytelling power and production cost, that's super cool and definitely counts as storyboards, but I suspect that's not the case.<p>[0]<a href="https://venturebeat.com/media/amazon-launches-storyteller-to-turn-scripts-into-storyboards-automagically/amp/" rel="nofollow">https://venturebeat.com/media/amazon-launches-storyteller-to...</a>