{"id":25375,"date":"2025-11-27T13:07:55","date_gmt":"2025-11-27T13:07:55","guid":{"rendered":"http:\/\/tezgyan.com\/index.php\/2025\/11\/27\/no-pun-no-gain-ai-is-only-pretending-to-understand-witty-jokes-tech-news\/"},"modified":"2025-11-27T13:07:55","modified_gmt":"2025-11-27T13:07:55","slug":"no-pun-no-gain-ai-is-only-pretending-to-understand-witty-jokes-tech-news","status":"publish","type":"post","link":"https:\/\/tezgyan.com\/index.php\/2025\/11\/27\/no-pun-no-gain-ai-is-only-pretending-to-understand-witty-jokes-tech-news\/","title":{"rendered":"No Pun, No Gain: AI Is Only Pretending To Understand Witty Jokes | Tech News"},"content":{"rendered":"


\n<\/p>\n

\n

Last Updated:<\/span>

It was earlier believed that AI processes punny humour in a way similar to humans. But a new study claims LLMs are only ‘pretending’ to understand wordplays and double-meanings<\/h2>\n
\n
\"It's<\/p>\n

It’s perhaps safe to say, from our study at least, that humans will always get the last laugh when it comes to comedy, says one of the study’s authors. (AI Generated)<\/p>\n<\/figure>\n<\/div>\n

Artificial Intelligence may be putting, or threatening to put, journalists, authors, screenwriters, actors and even singers out of jobs, but it is yet to touch the most basic form of entertainment \u2013 comedy. Yes, Large Language Models (LLMs) only pretend<\/em> to understand puns, wordplays and inuendoes, a new research has found. LLMs, like ChatGPT and Gemini, are powerful computer programs that learn from huge amounts of text so they can understand questions and respond in clear, human-like language.<\/p>\n

According to researchers at Cardiff University in Wales and Ca\u2019 Foscari University in Venice, LLMs or generative AI models are only able to \u201cspot the structure of a pun”, but don\u2019t really get the joke. It was earlier believed that AI processes such humour in a way similar to humans, but the new study argues that datasets at the time weren\u2019t meant to test pun interpretation by AI.<\/p>\n

How the researchers went about testing the LLMs now was by giving it two similar prompts \u2013 one with an obvious pun and another that replaced the pun with a synonym \u2013 to see whether it could tell the difference. What they found was that the AI misunderstood both sentences to have puns.<\/p>\n

Among the test sentences the researchers used was: \u201cI used to be a comedian, but my life became a joke.” They then tweaked it to say: \u201cI used to be a comedian, but my life became chaotic.”<\/p>\n

The LLMs erroneously thought both sentences contained puns, whereas only the first one does. (comedian, joke\u2026 get it?)<\/p>\n

Another example they reportedly tried was: \u201cLong fairy tales have a tendency to dragon.” (\u2018dragon\u2019 sounds like \u2018drag on\u2019). They then replaced the word \u2018dragon\u2019 with \u2018prolong\u2019, which means the same as \u2018drag on\u2019 but doesn\u2019t constitute any pun or funny wordplay. Again, the LLMs falsely concluded the presence of puns in both sentences.<\/p>\n

Replacing the pun with the synonym or similar-meaning words didn\u2019t change the meaning of the sentence on paper but obviously robbed it of humour, something the AI systems weren\u2019t able to detect.<\/p>\n

Perhaps the cheekiest one they tried was: \u201cOld LLMs never die, they just lose their attention.” The word \u2018attention\u2019 here serves a double meaning \u2013 it means \u2018concentration\u2019 as well as refers to a technical mechanism used in deep learning and data processing. When \u2018attention\u2019 was replaced with the contextually illogical word \u2018ukulele\u2019, the AI tools presented the forced reasoning that it could mean \u2018you-kill-LLM\u2019 that suggests the death of LLMs.<\/p>\n

The Guardian<\/em> quoted Cardiff University\u2019s Prof Jose Camacho Collados as saying that they were able to \u201cconsistently fool LLMs” by modifying puns.<\/p>\n

\u201cIn general, LLMs tend to memorise what they have learned in their training. As such, they catch existing puns well but that doesn\u2019t mean they truly understand them,” he said.<\/p>\n

\u201cIn these cases, models associate these sentences with previous puns, and make up all sort of reasons to justify they are a pun. Ultimately, we found their understanding of puns is an illusion,” he added.<\/p>\n

So what are the implications of these findings?<\/p>\n

\u201cIt\u2019s a reminder that, in general, outputs from these models should be taken with a pinch of salt,” TechXplore<\/em> quoted Camacho-Collados as saying. \u201cWhile AI is becoming more powerful, it\u2019s perhaps safe to say from our study at least, that humans will always get the last laugh when it comes to comedy.”<\/p>\n

\n
\n
<\/p>\n
\"Nitya<\/figure>\n

<\/span><\/p>\n

Nitya Thirumalai<\/a><\/p>\n

Nitya Thirumalai, News Editor at News18.com, writes on Indian and global politics as well as Formula 1. She was Google News Initiative-Columbia Journalism School Fellow in the inaugural Newsroom Leadership Prog…Read More<\/span><\/p>\n<\/div>\n<\/div>\n

Nitya Thirumalai, News Editor at News18.com, writes on Indian and global politics as well as Formula 1. She was Google News Initiative-Columbia Journalism School Fellow in the inaugural Newsroom Leadership Prog… Read More<\/span><\/p>\n<\/div>\n<\/div>\n

First Published:<\/span><\/p>\n
\n

November 27, 2025, 10:34 IST<\/p>\n<\/div>\n<\/div>\n

News<\/a> tech<\/a> No Pun, No Gain: AI Is Only Pretending To Understand Witty Jokes<\/span><\/div>\n
\n
\n
Disclaimer: Comments reflect users\u2019 views, not News18\u2019s. Please keep discussions respectful and constructive. Abusive, defamatory, or illegal comments will be removed. News18 may disable any comment at its discretion. By posting, you agree to our Terms of Use<\/a> and Privacy Policy<\/a>.<\/div>\n<\/div>\n<\/div>\n

Read More<\/span><\/p>\n<\/div>\n


\n
Source link <\/a><\/p>\n","protected":false},"excerpt":{"rendered":"

Last Updated:November 27, 2025, 12:18 IST It was earlier believed that AI processes punny humour in a way similar to humans. But a new study claims LLMs are only ‘pretending’ to understand wordplays and double-meanings It’s perhaps safe to say, from our study at least, that humans will always get the last laugh when it…<\/p>\n","protected":false},"author":1,"featured_media":25376,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[49],"tags":[],"class_list":["post-25375","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-tech"],"_links":{"self":[{"href":"https:\/\/tezgyan.com\/index.php\/wp-json\/wp\/v2\/posts\/25375","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/tezgyan.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/tezgyan.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/tezgyan.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/tezgyan.com\/index.php\/wp-json\/wp\/v2\/comments?post=25375"}],"version-history":[{"count":0,"href":"https:\/\/tezgyan.com\/index.php\/wp-json\/wp\/v2\/posts\/25375\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/tezgyan.com\/index.php\/wp-json\/wp\/v2\/media\/25376"}],"wp:attachment":[{"href":"https:\/\/tezgyan.com\/index.php\/wp-json\/wp\/v2\/media?parent=25375"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/tezgyan.com\/index.php\/wp-json\/wp\/v2\/categories?post=25375"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/tezgyan.com\/index.php\/wp-json\/wp\/v2\/tags?post=25375"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}