I think that sums up the problem with AI in general.
The real story so far of "AI" is about human desire to avoid responsibility for their actions. It is about saying "the magic box said you were a terrorist, nothing I can do about it" or, in this case, "The magic box said it should make a lot of money, I dunno what happened."
It is about the limited liability corporation gaining the ability to act semi-autonomously, with the profits going to the humans and the blame going to the machine.
You see the same in tech. It’s been a long time since the biggest players have done much innovative stuff.
I don't really watch movies myself, but I could tell you that Black Panther was a massive cultural success. As was Get Out. And The Shape Of Water (which I did see and enjoyed thoroughly, especially for its seamless integration with the battling of traditional and nontraditional masculinity). But I couldn't tell you anything about, I don't know, Antman or Daredevil other than that they existed at some point. And I think this reflects in commecial successes, also.
What has been forgotten is that masses do want some novelty, even if it isn't by very much. Incremental novelies, perhaps.
One of the biggest-grossing franchises in film history is rewarmed 30s SF pulp with a bigger effects budget.
>more awful movies while trying to argue that 'this is what people want.'
Moviegoers don't want the same old thing. But they also don't want this weird Lynch/Cronenberg (snobbish Godard) stuff. We should have books and films with enough cultural cachet that almost everyone should be experiencing them simultaneously (ie, there's too much fragmentation). But there should be small, independent, unique voices.
So when it comes right down to it, what we really want is something comfortable (known) that doesn't quite feel that way (but not too much) and everyone shares as a cultural touchstone, but one which is unique and special to each individual. This seems to be the real problem we're running against.
[0] https://www.nationalreview.com/2016/03/batman-v-superman-cul...
Even on ticket numbers for the big movies, everything is in decline. For instance Avengers Endgame is, by far, the biggest movie in recent decades. Yet it sold 3 million tickets fewer than e.g. Titanic. This is made even more remarkable by population growth. The population since 1998 (when Titanic was released) has increased by nearly 28%. Even the US population has increased by more than 19% since then. Titanic sold 94,524,324 tickets. A population normalized comparable success would move more than 112 million tickets, and that's using just a 19% growth factor.
Avengers: Endgame, the mega-hit of the 21st century, sold 91,353,296 tickets. The industry is dying and blaming absolutely everything except their own decisions which, in turn, is likely contributing to their own demise as they refuse to accept responsibility for their own declining results. One of the countless parallels shared between Hollywood and the big budget game industry.
Here's an extremely relevant article about Hollywood. [2] Hollywood has been one of the earliest adopters of active utilization and commercialization of "AI technologies". Scripts are actively analyzed using machine learning systems to try to predict profits. Guess what the AI says? Stuff that made money in the past will keep making money, and things that didn't won't. And what do we get out of Hollywood? Repetitive rubbish whose only novelty is tossing in some identity politics to, in this article's phrasing, offer a "more representative portrayal of the world in which we live."
More to the point, think about some of the all time greatest movies (as well as money makers). Here's [3] a list of the top 10 by gross. Tellingly, only 3 of the top 10 were made after 2000. But now consider the other films - Gone with the Wind, Titanic, Star Wars, The Sound of Music, ET, The Ten Commandments, etc. These films were all, if not extremely unique, certainly not derivative rubbish. Now look at the 21st century films and you'll see a whole bunch of generic superhero movies and sequels. Even if not the product of AI, it's certainly the product of a mindset that's no different than a profit oriented regression analysis of the past - the less glorified truism of what "AI analysis" is.
Maybe the biggest problem with Hollywood is summed by one sentence in the AI article I referenced: “[AI analysis] takes a lot of the risk out of what I do.” No, it just makes you comfortable releasing crap that you'd otherwise recognize as crap. Interestingly the producer that gave that quote ended up releasing two mega-flops after the article and has not released anything since. It's not like we've even scratched the surface of possible ideas, but Hollywood has become uncreative, lazy, and averse to any and all risk - to the point that they'd rather drown than try something else, so long as it happens slowly.
[1] - https://www.the-numbers.com/market/
[2] - https://www.nytimes.com/2013/05/06/business/media/solving-eq...
[3] - https://en.wikipedia.org/wiki/List_of_highest-grossing_films...