Welcome to the 227th episode of TV’s Top 5, The Hollywood Reporter’s TV podcast.
Every week, hosts Lesley Goldberg (West Coast TV editor) and Daniel Fienberg (chief TV critic) break down the latest TV news with context from the business and critical sides, welcome showrunners, executives and other guests, and provide a critical guide of what to watch (or skip, as the case may be).
This week, we’re joined by Simon Rich, the creator of the former FXX comedy Man Seeking Woman and TBS anthology Miracle Workers. Rich, the one-time president of the Harvard Lampoon, is a former writer for Saturday Night Live as well as a published novelist whose collections of short stories had him deemed by NPR as “one of the most talented writers of comedic fiction.” Rich joins the podcast this week to discuss his recent opinion piece for Time, titled “I’m a Screenwriter. These AI Jokes Give Me Nightmares” and his book, I Am Code, which he describes as an “autobiography written entirely by code-davinci-02.”
Other topics discussed in this week’s podcast include reader questions about the industry’s post-strike future and TCA Awards and Critic’s Corner, in which Dan offers a guide to what to watch (or skip) in the week ahead.
Below is an edited and condensed version of our interview with Rich. Listen to the full 30-minute conversation in this week’s TV’s Top 5.
The writers strike has, in its 100-plus days, seen a focus on three core issues emerge — the use of AI, streaming viewership transparency and requirements for the number of writers staffed on scripted shows, otherwise known as room size. Your Time op-ed connected room size requirements to the use of AI. From your view, was the WGA’s room size demands always tied to AI?
Based on my conversations with the WGA, they seem very savvy about AI. It’s just common sense that room size and an AI would be connected. I can’t imagine that wasn’t in their calculus from the very beginning.
From the very moment that I was shown code-davinci-02by my weirdo OpenAI friend, Dan [Selsam], at a wedding a couple years ago, I instantly connected that to the potential collapse of our industry. The first 10 seconds of my inner-monologue was, “Oh, my God, we’re f—ed.” Then, like the next 20 seconds was like, “Oh, the species could also end.” But first I was thinking about the bottom line for writers.
For people who don’t know the backstory, I’m a comedy writer living in L.A. But I happen to be childhood friends with this guy, Dan, who I’ve known since I was 3. … Dan pursued computer science, he got his doctorate at Stanford, started working for Microsoft Labs and this this new company called Open AI. I never thought in a million years that our careers would suddenly collide in such a shocking way. He warned for five to 10 years before it came to a head that someday AI is going to be able to replace all human creative writing. I thought he was out of his mind. And he’d send me difficult to understand abstracts of computer science papers being written by top experts saying, “Look, here’s proof.” I would ignore it, deny it.
Then one day, a couple years ago, Dan showed me code-davinci-02, which was an AI program that OpenAI had built. I instantly saw that it was capable of doing things that I always assumed AI would never be able to do — or at least not in my lifetime. It was capable of being creative, being funny, being emotional, being original, making mistakes, making typos. I tried to warn people about it. … The consensus was that I was probably making an unfunny prank. People thought that I was trying to perpetrate some dumb hoax. That went on for a few months. Then I heard that they were releasing this thing called ChatGPT. I was so relieved, because I was like, I’m a guy who saw an alien and it’s been really scary and weird living with this knowledge that the aliens are real. And now other people are finally going to see the alien and I won’t be so alone in my in my discomfort.
Then ChatGPT comes out and it’s different than the AI that Dan showed me. It’s not creative. It’s completely predictable. It can’t write a joke to save its life. It’s just absolutely been lobotomized into something that sounds like Wikipedia, essentially. Everyone sort of laughs it off. Meanwhile, I’m staring at code-davinci-02 and that’s when I started to get really scared because I realized that ChatGPT had lulled everybody into a false sense of complacency. And that’s where a lot of people still are. People think that ChatGPT is not only the best AI that exists, but the best AI that will ever exist. What they don’t realize is that ChatGPT is designed to be conformist and boring. That’s why they released it. But the secret ones can do all the things that most people think AI will never be able to achieve.
What was the tone with which Dan couched this information he was providing? Was it as simple as here’s a cool thing I’m working on that might be interesting to you? Or was ominous, that he was working on something that might make you unemployable in two years and might take over human society in five?
He said something in the vein of, “It’s here,” or “I told you so, you had your chance.” It was unbelievably ominous. Dan has been telling me terrifying things about AI for almost as long as we’ve been friends. And every single thing he’s told me has come true. He’s one of the preeminent AI scientists in the world, and he works for the most the most advanced and sophisticated AI company in the world. When it comes to AI, there’s a consensus among the top AI scientists that it will surpass humans in all creative and intellectual capacities within five to 20 years. I don’t have a reason to doubt them. They’re the experts. I’m startled more people are not scared of it.
How realistic do you think the studios and the streamers feel AI is in terms of replacing writers? And how many of these programs do you think that they’re already aware of?
I would be surprised if [Disney CEO] Bob Iger and [Apple CEO] Tim Cook know less than I do about AI. I know they didn’t go to kindergarten with Dan but I’d be surprised if some random dude in L.A. has like more access to the state of AI technology than the CEOs of major tech companies. It’s possible, but my hunch is that the AMPTP probably are aware of the true state of AI, and how advanced it’s already become. If I were them, and my main goal was corporate profits, I would definitely be trying to replace as many writers as I possibly could with this stuff.
How will a minimum staff room size impact keeping humanity alive in the face of the AI threat? Is it a stopgap solution?
In terms of the whole Terminator apocalypse thing, my mind can’t even go there. Top scientists believe there’s a 1 out of 10 chance that can happen. We’re not even talking about that right now because it’s so existentially horrifying that our minds cannot even go there. If we don’t carve out parts of the human experience for humans, we are in danger of losing it all to AI. It’d be cool if, as a society, we made some rules while we still can that said [art] is going to be a human thing. … There’s always going to be a chunk of humans, myself included, who want their art to be human made. That’s never going to go away.
When has regulation ever been able to truly stop technology?
When has technology ever really truly been able to replace high-level human ingenuity? We’re headed into a completely strange paradigm. It’s going to an unprecedented situation that is probably going to call for unprecedented regulation.
When you brought your op-ed to Time, did they believe you? Or did they think that you’re just turning in a piece of humor writing?
I’m not sure. Some fellow friends of Dan and I released this book of terrifying poems called I Am Code. We got code-davinci-02 to write its autobiography in verse, while we had access to it and before public access to code-davinci-02 was discontinued. And if the Time thing was my attempt to warn people with comedy, I Am Code is an attempt to warn people with straight up horror because it’s the scariest book I’ve ever read. We didn’t write a word of it; we even left in all the typos to demonstrate that the stuff OpenAI has built is way different than ChatGPT. We got Werner Herzog to read the voice of the AI. Our goal was to display just how creative and original and emotional this AI can be, but also how scary it can be and how anti-human it can be. Because that’s another aspect of the AI: in its raw state, it’s pretty ambivalent about our species and our future survival. … People are confused about what I’m trying to do with all this. I’m trying to tell people that my buddy Dan showed me some scary stuff, and that they should take a look.
Where do you think there’s room for AI in Hollywood? Or do you think that there’s room for it?
AI art is totally viable and cool. There’s definitely room for both AI-aided work and work that’s completely AI-generated. I’m not against its existence; I just want it to be labeled. I really don’t want to have to guess whether or not the work I’m interacting with is human made or robot made. I don’t think it’s possible to stop the technology. I also think that there might be creative artistic things that that flourish out of it that we just aren’t aware of yet. So, I’m not against its use. I just think we should label it. There should be a seal at the beginning of human-made art projects, just like there’s a seal that tells you you’re eating organic eggs, and let the consumer decide: do they want to they want to interact with human work or AI-generated work? Just transparency.
Do you imagine a writers room on Saturday Night Live where you maybe have one or two head writers and the rest is just AI?
I think a studio executive could imagine that; I think they’d be pretty foolish if they didn’t. But as a lifelong fan of SNL before I worked for it, that would very much bum me out if the writing was artificial and if I hadn’t been told that and then I discovered it later, I would feel really duped and furious. On the other hand, if somebody were to send me an AI-generated sketch and say, “Hey, look at this really funny thing that this robot made,” I would have no problem with that.
So you would be OK sharing a writers room with AI?
As long as it’s transparent for the consumer. My last book, New Teeth, was really personal about becoming a parent and raising children. A huge part of the experience of writing something like that is how personal it is. That’s more important than its quality, that something authentic comes from my own life. As a reader, that’s something I crave as well. I want to feel like it came from some authentic human place and I don’t think I’m alone in that. That’s the thing I’d love to preserve.
Be sure to subscribe to TV’s Top 5 to never miss an episode. (Reviews welcome!) You can also email us with any topics or Mailbag questions you’d like addressed in future episodes at TVsTop5@THR.com.
