There's one thing AI can't do: be funny
But will the technology's hilarity evolve? Some experts think so.
Artificial intelligence is prolific. It can write academic essays, diagnose diseases and power self-driving cars. But lucky for those who fear the eventual dominion of robot overlords, there is one area where artificial intelligence falls flat: humor. This should hearten comedy writers, who worry AI may take their already-scarce jobs. "ChatGPT Doesn't Have Childhood Trauma," said the sign of one writer at the Writers Guild strikes last year, a rallying cry for those who do not believe AI can create meaningful art.
If you ask AI to be funny, it spits out clichés. "According to one 2023 study in which researchers asked for 1,008 'original' jokes from ChatGPT, more than 90% of replies were the same 25 jokes … none of which were original," said Mashable. AI can be funny unintentionally — "like the oblivious 'straight man' in sketch comedy" — but that is something else entirely. So why, with an endless breadth of knowledge at its disposal, is artificial intelligence seemingly incapable of making us laugh?
Why is AI unable to be funny?
Comedy is situational, contextual and sophisticated. Even a simple meme can require several layers of understanding to "get." While receiving her Ph.D. at Delaware University in 2021, Ishaani Priyadarshini wrote her dissertation about AI's apparent inability to decipher memes, according to Popular Mechanics. "Internet memes serve as excellent checkpoints to ensure humans have the upper hand over machines," Priyadarshini said to the outlet.
Subscribe to The Week
Escape your echo chamber. Get the facts behind the news, plus analysis from multiple perspectives.
Sign up for The Week's Free Newsletters
From our morning news briefing to a weekly Good News Newsletter, get the best of The Week delivered directly to your inbox.
From our morning news briefing to a weekly Good News Newsletter, get the best of The Week delivered directly to your inbox.
Humor is subjective, but AI has no subjectivity or personal feelings about anything. The best comedy is littered with emotional truths, yet as the Writers Guild sign noted, AI has no childhood trauma, no well of empathy or lived experiences to pull from. "Large language models can be taught to whip up passable formulaic material because the propagation of hack jokes relies on systematized pattern-recognition processes," said The Hollywood Reporter. "However, [experts] believe that original, path-breaking comedy will likely remain out of the conceptual reach of such machinery, at least in the near term."
Another challenge: Many comedy routines push the boundaries of propriety and political correctness, while "AI, a conservative technology, doesn't understand what taboos are, so it can't break them," said Guy Hoffman, a Cornell professor, to the Reporter. AI struggles to defy established norms because it has been programmed to follow them.
Will AI ever evolve to be funny?
Some experts believe AI could advance. A robot comedian named Jon went on a stand-up tour as part of a study by Oregon State University in 2020, though humans wrote and programmed his set list. It is "possible to evolve humor in something like the ChatGPT model by programming it to favor the incongruities and deviations from established norms that are the hallmarks of comedy," said Tony Veale, a professor at University College Dublin and the author of "Your Wit Is My Command: Building AIs With a Sense of Humor," to the Reporter.
"A language model uses probability to guide its choices," Veale said to the outlet. "You could adjust its probability controls, from the expected to the unexpected, since the dominant theories of humor have to do with taking ideas and subverting them."
AI has not been given much time to hone its craft, something most comedians do at open mics for years before finding success. Still, some experts say humans maintain a clear advantage. "It's a little bit like an alien watching a standup routine and saying 'oh, I can do that, I think I understand the pattern,'" Mark Riedl, a professor at Georgia Tech's Machine Learning Center, said to Mashable. "But [the alien] doesn't realize there's a lot more going on in the mind of the comedian and the minds of the audience."
Trust is another factor. "Studies have found that people rate humor as one of the tasks they trust humans with far more than AI, along with writing news articles, composing songs and driving trucks (all of which AI has some success in doing)," said Time Magazine.
The evolution of AI's hilarity could spell trouble, as its mediocre quips may eventually become passable enough to steal some comedians' jobs. "The consensus among experts is that AI, while a clear labor threat, will become a baseline tool for comedy writers, like a thesaurus or a search engine," said the Reporter.
Sign up for Today's Best Articles in your inbox
A free daily email with the biggest news stories of the day – and the best features from TheWeek.com
Anya Jaremko-Greenwold has worked as a story editor at The Week since 2024. She previously worked at FLOOD Magazine, Woman's World, First for Women, DGO Magazine and BOMB Magazine. Anya's culture writing has appeared in The Atlantic, Jezebel, Vice and the Los Angeles Review of Books, among others.
-
'It may not be surprising that creative work is used without permission'
Instant Opinion Opinion, comment and editorials of the day
By Justin Klawans, The Week US Published
-
5 simple items to help make your airplane seat more comfortable
The Week Recommends Gel cushions and inflatable travel pillows make a world of difference
By Catherine Garcia, The Week US Published
-
How safe are cruise ships in storms?
The Explainer The vessels are always prepared
By Devika Rao, The Week US Published
-
What Trump's win could mean for Big Tech
Talking Points The tech industry is bracing itself for Trump's second administration
By Theara Coleman, The Week US Published
-
Google Maps gets an AI upgrade to compete with Apple
Under the Radar The Google-owned Waze, a navigation app, will be getting similar upgrades
By Justin Klawans, The Week US Published
-
Australia proposes social media ban before age 16
Speed Read Australia proposes social media ban before age 16
By Peter Weber, The Week US Published
-
Is ChatGPT's new search engine OpenAI's Google 'killer'?
Talking Point There's a new AI-backed search engine in town. But can it stand up to Google's decades-long hold on internet searches?
By Theara Coleman, The Week US Published
-
Teen suicide puts AI chatbots in the hot seat
In the spotlight A Florida mom has targeted custom AI chatbot platform Character.AI and Google in a lawsuit over her son's death
By Theara Coleman, The Week US Published
-
FTC bans fake online product reviews
Speed Read The agency will enforce fines of up to $51,744 per violation
By Peter Weber, The Week US Published
-
The internet is being overrun by ads
Under the Radar Grabbing attention has never been more annoying
By Devika Rao, The Week US Published
-
States sue TikTok over children's mental health
Speed Read The lawsuit was filed by 13 states and Washington, D.C.
By Rafi Schwartz, The Week US Published