Adobe Stock
Many comedy writers regard artificial intelligence as a no laughing matter. That’s awful, Seth Rogen explained to The Hollywood Reporter at the May 10 premiere of the Apple TV+ show Platonic, in which he plays. Any use of AI seems terrifying and also just plain financially unfair because it’s all injected with stuff they don’t keep track of, referring to how the technology is trained on material without the consent of its creators.
The distrust intensified during the ongoing writers’ strike. A sticking point in the failed contract negotiations between the WGA and the AMPTP, which negotiates on behalf of the studios, is a proposal to regulate the use of artificial intelligence. The machines would not be allowed to write or rewrite literary material, or be used as a source, nor could the syndicated output be used to train AI models. The studios rejected that plan, instead offering to meet only once a year to discuss issues presented by wayward technology.
Writers Guild members see AI not only as a looming threat, but also as a clear and present danger, believing that the business landscape will inexorably be changed by the next negotiating round several years from now. ChatGPT has no childhood trauma, read a sign at the Radford Studio Center in the San Fernando Valley. A robot cannot feel shame, explained Harris Meyersohn (Tha Gods Honest Truth with Charlamagne Tha God) on the first day of protests outside Netflix’s Hollywood headquarters. It’s essential to be human and essential to be a writer.
For the most part, scientists in the field of artificial intelligence research known as computational humor think the immediate concern is overblown. These experts, some of whom have studied funny robot issues for decades, note that large speech patterns can be taught to produce passable formulaic material because the propagation of hack jokes relies on systematized pattern recognition processes. However, they believe that original and groundbreaking comedy is likely to remain beyond the conceptual reach of such machinery, at least in the short term, because it is such a singularly complex and silvery language. Humor is highly contextual and situational, making it an extraordinarily difficult problem to solve, explains Mark Riedl of Georgia Tech.
These specialists argue that AI appears to address two essential challenges in crafting top-notch comedy. (While newly released image and music generators have wowed users with results that often go viral, the latest wave of chatbots has thus far produced no equivalent moment of uproarious laughter.) While it has access to unfathomable knowledge, it can only exploit to approximate life experience. On a fundamental level, he can always have difficulty understanding humans. Unless the machine understands Why a joke is funny, you’re nowhere, says Julia Rayz of Purdue. DARPA, the Defense Department agency that helped pioneer the Internet, could help remedy that. Now he is trying to teach computers about cross-cultural communication by developing natural language processing technologies.
The other problem is that AI, which synthesizes existing datasets, would appear to be at a disadvantage if the goal is to generate edgy, boundary-pushing results: innovative gags, wayward assumptions, unpredictable tonal decisions. Artificial intelligence, a conservative technology, doesn’t understand what taboos are, so it can’t break them, notes Cornell’s Guy Hoffman.
Not all are naysayers. Tony Veale of University College Dublin, who explains that being (intentionally) funny requires the cognitive ability to understand and attribute mental states to oneself and others, what is known as theory of mind, thinks it is possible to evolve l There is humor in something like the ChatGPT model by programming it to favor the inconsistencies and deviations from established norms that are the hallmarks of comedy. A language model uses probability to guide its choices, says Veale, the author of Your wit is my command: build AI with a sense of humor. You could adjust its probability controls, from the expected to the unexpected, as the dominant theories of humor have to do with taking ideas and subverting them.
Joe Toplyn, former head writer for both Jay Leno and David Letterman, has already built his own joke-making bot, Witscript, based on joke algorithms he previously developed, patented, and taught at the Peoples Improv Theater in Manhattan. When it works best, he writes jokes good enough to be used on a late-night comedy talk show, without any editing, he says. Toplyn, a WGA member with an engineering degree from Harvard who recently picketed outside the HBO office in Manhattan while holding up a sign that reads Don’t Let ChatGPT Write Yellowstone, insists the future is already here. She explains that she developed his beta software for him, which he demonstrated DAY, to help non-professionals who can’t write jokes, but thinks if producers had access to them, they could cut costs for writers. (As part of a February part, Jimmy Kimmel rehearsed AI-generated jokes, one of which earned a smattering of genuine laughs from his audience.)
The commercial issue, as always, is quality. If brilliance is the business plan, then the human humorist is safe. But if not, the AI, even if it achieves just the bare minimum of hilarity, could be just fine. How mediocre do you agree with your being a comedian? muses Piotr Mirowski, a scientist employed by a machine learning company who also co-founded an AI-enabled improv company that incorporates a chatbot into its performances. It is up to us to judge the results of AI with discernment.
The consensus among these experts is that AI, while a clear threat to work, will become a basic tool for comedy writers, like a thesaurus or search engine. Diyi Yang at Stanford notes that, in researching her, professional stand-up comedians may not have found the AI-generated jokes amusing. Yet often inspired by the patterns of strange and unconventional verbal associations, this thrill is similar to musician Brian Enos’ use of Oblique Strategy Cards, which are intended to break creative blocks and promote unconventional thinking.
As Columbia’s Tuhin Chakrabarty says of AI and humor, there must always be humans around. Adds Simon Colton of Queen Mary University of London: It will still make commercial sense to hire a professional. Good companies will realize it’s a multiplier.
For his part, Jeff Schaffer, well-known comedy writer (Curb your enthusiasm) and creator (Dave), declares that he doesn’t care about the increase in AI. I don’t know much about AI, but I know something about people, she says. So far, AI is no fun. And that’s his most human quality because most people aren’t even funny. And the more AI becomes like people, the more it will think It is funny. And so if it becomes exactly like people: the AI will think it’s funny, and everyone else will hear or read it and say, “The kid thinks he is.” AS funny but it’s so boring.’
Jackie Strause and Kirsten Chuba contributed to this report.
A version of this story first appeared in the May 31 issue of The Hollywood Reporter magazine. Click here to register now.
#isnt #fun