Title: Is ChatGPT funnier than real humans?
An intriguing study conducted by researchers at the University of Southern California (USC) sought to answer the age-old question: Can a machine be funnier than a human? The team compared the humor capabilities of ChatGPT 3.5, a popular large language model, with a group of U.S. adults in two separate studies.
In the first study, participants were tasked with a comedy writing assignment, and their responses were anonymously rated by another group of adults. Surprisingly, the majority (69.5 percent) found ChatGPT’s jokes funnier compared to human-generated responses (26.5 percent). This suggests that the AI-powered model has a knack for humor that rivals that of humans, with 4 percent finding both equally humorous.
Moving on to the second study, ChatGPT was challenged with creating satirical headlines inspired by The Onion, a renowned source of satire. The new headlines were then presented to a group of USC psychology students for rating. While most preferred The Onion’s headlines (48.8 percent), ChatGPT’s headlines were not far behind at 36.9 percent. This indicates that the AI model can produce humor comparable to that of professional comedy writers.
Lead author Drew Gorenz, a doctoral candidate in social psychology at USC, highlighted the significance of these findings. Despite ChatGPT’s inability to experience emotions, its ability to generate original and humorous content demonstrates that emotional connection is not a prerequisite for creating top-notch jokes.
In conclusion, the studies shed light on the evolving landscape of humor creation, where AI models like ChatGPT are proving to be formidable contenders against human comedians. While humor remains subjective, the research underscores the AI model’s potential to craft witty and entertaining content that resonates with audiences.
