Skip to main content
Go to the home page of the European Commission (opens in new window)
English English
CORDIS - EU research results
CORDIS

Article Category

Article available in the following languages:

AI and humour, is it a laughing matter?

AI may be able to generate jokes, but does it understand them?

AI has many eye-opening capabilities, from diagnosing tumours and writing poetry to passing academic exams. But will it ever grasp the nuances of humour – one of humanity’s key traits? According to a paper published in the ‘Association for Computational Linguistics’(opens in new window), AI doesn’t understand what makes jokes funny. A research team at Cornell University in the United States tested AI models and humans on tasks that involved hundreds of entries from The New Yorker’s Cartoon Caption Contest. Participants were asked to match jokes to cartoons, identify winning captions and explain why a winning caption is funny.

Human vs AI? Competition is laughable

The results showed that humans greatly outperformed the machines. AI models attained 62 % accuracy in a multiple-choice test that matched cartoons to captions, compared with 94 % by humans. In contrasting human versus AI-generated explanations of humour, human versions were favoured about 2 to 1. However, the study indicated that AI could still be a valuable tool for humourists who brainstorm ideas. “The way people challenge AI models for understanding is to build tests for them – multiple choice tests or other evaluations with an accuracy score,” explained Dr Jack Hessel, a research scientist at the non-profit Allen Institute for AI, in a Cornell University news item(opens in new window). “And if a model eventually surpasses whatever humans get at this test, you think, ‘OK, does this mean it truly understands?’ It’s a defensible position to say that no machine can truly ‘understand’ because understanding is a human thing. But, whether the machine understands or not, it’s still impressive how well they do on these tasks.” “There are datasets of photos from Flickr with captions like, ‘This is my dog’,” he elaborated further. “The interesting thing about the New Yorker case is that the relationships between the images and the captions are indirect, playful, and reference lots of real-world entities and norms. And so the task of ‘understanding’ the relationship between these things requires a bit more sophistication.”

Who will get the last laugh?

With AI’s potential to upend professions, should comedians fear for their livelihoods? “[W]here true performers and people with actual charisma and comedic wherewithal will always flourish is no one’s going to show up to a comedy club and buy a two-drink minimum to stare at a laptop, typing out words, or even saying those words through some Bluetooth audio,” popular American stand-up comic Jimmy Failla told ‘Fox News’(opens in new window). “So I don’t necessarily feel threatened by it.” He added: “People laugh at the jokes from humans because they can feel the humanity in them. There is a shared truth. I can promise you the artificial intelligence comic isn’t going to have that shared experience.”

My booklet 0 0