Study: ChatGPT is funnier than humans

Updated July 8, 2024, 4:32 a.m. | By East Coast Breakfast

Count your days, comedians, because AI-generated jokes could soon threaten your jobs...

small artificial intelligence assistant robot — File photo: iStock/PhonlamaiPhoto (modified)

A new study by the University of Southern California (USC) has found that jokes generated by ChatGPT are more hilarious than those written by humans.

The study, conducted by USC psychology, was published on July 3. It involved two tests. The first test compared AI-generated jokes on OpenAI's ChatGPT 3.5 to those created by laypeople - those without professional comedy writing experience.

Participants were asked to complete three different humour tasks and various prompts.

The results found that 69.5% of people who compared the jokes rated the AI responses as funnier than the human responses. Just 4% found the AI-generated jokes and human-created jokes equally funny.

"Despite ChatGPT’s inability to “feel” emotions, it outperformed the average human participant across all three humour production tasks. Overall, ChatGPT 3.5 performed above 63% to 87% of human participants depending on the humour test. ChatGPT 3.5 showed particularly strong performance in the roast joke task. We found this result particularly interesting given the aggressive nature of the task. Given that ChatGPT is designed not to generate any speech that could be considered offensive or hateful, the opposite prediction could have been made," the study stated.

ALSO READ: ChatGPT gets eyes: Free AI will now understand our world through video

The second part of the study compared AI-generated jokes on ChatGPT 3.5 to jokes written by professional writers at The Onion - a popular American publication known for its satirical news articles.

In one task, ChatGPT was asked to rewrite some of The Onion's headlines. So, how did the AI-generated satirical headlines compare to those created by The Onion’s team of professional comedic writers?

Pretty good! In fact, many people found them just as funny.

"In sum, ChatGPT produced novel, satirical news headlines of comparable quality to The Onion. Participants, on average, rated the headlines as similarly funny, indicating that the average participant did not discern a difference in quality. This is particularly interesting given the high standard of comparison (i.e., professional comedy writers) in this study," the study found.

MORE FROM EAST COAST RADIO:

This could spell bad news for comedians and professional comedic writers.

The study says there will be "large economic implications" if large language models (LLMs) like ChatGPT can generate comparable output to professional comedians and writers.

"Future research should investigate the potential of using LLMs for comedic writing across other commercially successful formats such as script writing, cartoon captioning, and meme generation once LLMs begin to incorporate image generation capabilities."

Check out the full USC study in the journal PLOS ONE.

ALSO READ: Say what now? Hobby horsing is a real thing and has its own championship