It's the summer of strikes in the entertainment industry.
The Writers Guild struck on May 2nd, with the Screen Actor's Guild following on July 14th. Both unions shared many of the same complaints - shortened seasons causing less pay for the actors and writers, the absolute dearth of residuals from streaming services, and the use of artificial intelligence.
There's been a surge of interest in the topic lately. Much of it was kickstarted by the introduction of ChatGPT late last year. ChatGPT is a large language model chatbot that has been getting scarily good at mimicking human writing. Coupled with the surge in AI generated art, there's a real feeling that this might mark a significant step forward in artificial intelligence. (At least once the AI figures out how many fingers humans have.) Still, things have progressed far enough that the new Marvel show on Disney+, Secret Invasion, used AI to generate the opening credits.
The WGA has demanded that no AI be used to write or rewrite scripts. Now, after having seen some attempts to get AI to do writing based on prompts, you might be saying "What's the big deal? Those writing prompts were terrible." And that is true! G/O Media tried to experiment with some posts on their websites that were written by AI, and it was a disaster. The AI wrote a post for their science fiction site, io9.com, that was listing the Star Wars movies in order of their release and it got it wrong. This is something that you could find on any number of websites, and the AI could not scrape it together. (Just as an aside, AV Club, io9, and Deadspin were some of my all-time favorite sites to read and comment on, and G/O Media has utterly destroyed them over the last several years. G/O Media can go right to hell.)
So if the AI cannot look up something like "Star Wars films release dates" on Wikipedia or IMDB, then how could it possibly write something truly original and creative?
Here's the thing.
I am not worried about AI or ChatGPT writing a masterpiece.
I am worried about them becoming mediocre.
I do not think an AI is going to write the next Slaughterhouse-Five or Eternal Sunshine of the Spotless Mind or Everything Everywhere All At Once. AI bots scrape the web for existing work and add it to their database. No AI is going to create a personal, inventive, or heartfelt original story.
I do think that AI may get to a point where it can spit out ten episodes of a cliched, cop show pastiche.
People watch shows differently that they used to. There are very few shows that demand your rapt attention, like Game of Thrones or Succession, where you'll miss a huge plot reveal if you look away to scroll through Twitter (I refuse to call it X.) or play a round of Marvel Snap. (God, that game is like crack.)
TV is just on for a lot of people. Goodness knows, I've done that. I like having Law & Order on in the background while I do chores or write. I don't have to catch every word, since I know the plots are basically on rails. I just like hearing Jerry Orbach and Sam Waterson. Also, it's the only reason I've ever seen Bones. I've never once in my life said "Hey, you know what I want to watch? Bones! That show with the improbably hot scientists." Yet I have seen dozens of episodes because it came on after Law & Order on TNT for a while and it was just okay enough of a show that I didn't change the channel.
So what if instead of Bones, TNT put on a procedural "written" by an AI? Suppose the AI scans enough scripts from decades of cop shows that they can put the cliches together in the right order. Maybe the show has two mismatched partners. Maybe one is a gung ho rookie, paired with a jaded veteran. Maybe one is a straight arrow, and the other plays by his own rules. And they have a hard ass lieutenant who gives them 24 hours to crack the case. If you weren't paying close attention, would you even notice or care?
I probably wouldn't. If I was deep in writing a chapter, I might poke my head up when the new theme song started but I probably wouldn't care enough to change the channel.
But the networks and producers care! If the bar is "just ok enough that I don't instantly change the channel," that's low enough for them. They're quite happy to pump out garbage if it means lower costs and fewer residuals to pay out.
Now, you may ask “so what?” If a show is bad or forgettable, who cares if it was written by a human or a computer?
Well, you should, for one. You know how TV writers learn their trade? By working on shows, and they can’t all be The Wire. They break in, working in the writing rooms and then they get hired on other shows. George RR Martin wrote about his experiences starting off as incredibly green writer on the Twilight Zone in the ‘80s. That was how he learned how to make TV shows. And if computers write the crappy shows, humans aren’t going to get the experience they need to write the good ones.
Netflix and Amazon produce an incredible amount of content, most of which you've never heard of. NPR pop culture writer Linda Holmes asked her Twitter followers if anyone realized that a remake of the 1994 Meryl Streep movie The River Wild was being released, now starring Leighton Meester and Adam Brody. Only about 7% had heard about it. That's a film based on a recognizable title with actors you've seen in things. It's not something they dredged up from the depths of Tubi. If the streamers just need CONTENT, why bother with making it good? Why bother advertising? Cut your costs, make the shareholders happy, and get your golden parachute packed.
In the battle of art vs. commerce, commerce will usually win. The streamers and studios have the money and don't want to share it. They will be perfectly happy to have one intern scan scripts into an AI and produce dreck rather than fairly compensate the writers. They will count on us all being too passive and distracted to notice.
Maybe they're right.
But I take solace in Barbenheimer.
Audiences have flocked to Chris Nolan's three hour biopic about the creator of the atomic bomb and the sly comedy that Greta Gerwig made out of what should have been - by all rights - a two hour toy commercial. Both are very personal and creative takes that could have in no way been created by an AI that read a Wikipedia article.
Barbie is on track to make over a billion dollars and Oppenheimer is on pace for about half that. There is money to be made if you are willing to let smart filmmakers do something creative.
Will studios understand that? Or will they instead try and write a Hungry Hungry Hippos movie with an AI?
We'll find out soon.
Victor Catano lives in New York City with his wonderful wife, Kim, and his adorable pughuaua, Danerys. When not writing, he works in live theater as a stage manager, production manager, and chaos coordinator. His hobbies include coffee, Broadway musicals, and complaining about the NY Mets and Philadelphia Eagles. Follow him on Twitter and Instagram @vgcatano and find his books on Amazon
No comments:
Post a Comment