The Complexity of A.I.

I just finished an historically accurate Political Noir feature film screenplay, Boyle Heights 1939.

I used A.I. for research, a LOT of research, into 1939.

It went deeper and faster in research than I ever could, and I have always done my own research (FUN FACT: I was a Research Associate at UCLA Medical School).

15 years ago, I wrote an historical fiction novel set in 1915 that took me several weeks and travel to libraries, etc. to get half the information that A.I. uncovered for this screenplay in days of research.

By the way, I learned then that not everything was on the Internet. I had to go to stacks at libraries because I was researching Jim Crow segregation and, surprise! (not) a lot of Black authors and subjects were still hidden in print books in library stacks, not apparently worthy of being ingested by the ‘Net.

The depth of research now, via A.I., was fantastic, and very helpful facts I might not have noticed quickly popped up.

This allowed me to build a rich, accurate world for my story, which is very, very important. My story is about the prominant NAZIs and antisemitism in America before World War Two and how that is being replicated today. That's a strong and controversial premise, one that will receive blowback by “half” of America. That half will destroy the entire premise of the story if they can attack one faulty fact (“There was no Hills Brothers ground coffee in 1939, so that proves the entire story is bullshit!”) NOTE: Yes, there was, and I have an unopened can of 1930s Hills Brothers coffee.

So, it is quite a burden to be factual, especially in an era of fact-free politics.

But just as with all critical thinking skills, I used four and five different A.I. tools to assure that all the information was consistent.

Guess what? A.I. made the same mistakes that humans do! I caught mistakes because I used multiple valid sources for verification, which is known as the lost art of “critical thinking skills.”

Unlike humans, though, when I called out the mistakes, A.I. just made corrections without a hissy fit.

I have a folder of Historical Research that has 250 files in it of research, each file 5 to 50 pages. Many are A.I. derived, many are from books, newspapers, and other sources.

Since I am all alone, not by choice but because of ageism, I also brainstormed with A.I., just as I used to do with fellow writers. A.I. is good for brainstorming.

A.I. rarely understood the emotions I ascribed to the characters and audiences, and even though it made a lot of suggestions, I rejected almost all of them. In the entire screenplay, there are two, maybe three sentences that were inspired by A.I. … just as if, had a human being said those sentences, I would have included them. They were helpful for the story.

One major "flaw" of A.I. is that it wants to flatter you so you keep interacting with it. I had to add "Please be straight with your response and do not flatter me" to every prompt. I urge EVERYBODY who uses A.I. to do this.

Here’s the thing:

If we are scared of all A.I., I don't know where we will end up, because it is here and it is a tool. It has now outgrown the Internet (I used to declare that the only way to stop A.I. would be to turn off the World Wide Web, but that is no longer true).

Absolutely, A.I. needs to be tamed. It will transform jobs and eliminate jobs (just like MAGA David Ellison will do when combining his empire, the Paramount empire, and the Warner Bros. empire into a single state propaganda machine, regardless of A.I.)

The fact is, we are in the new industrial revolution, and it's coming at us faster than Internet speed. We can't stay down on the farm and we've got to move … fast.