ChatGPT Seems to Like to Tell Unprompted 9/11 Jokes

Just when you thought only Elon Musk's image and chat generator Grok was the only thing capable of over-the-top crass imagery and definitely-not-PC jokes, a new competitor in the field of artificially intelligent comedians has joined the stage: ChatGPT. 

Developed by OpenAI and Microsoft in 2022, ChatGPT has taken the world by storm — for better or for worse — making tedious workplace emails faster, but also causing teachers massive headaches. The bot can write short stories, create detailed travel itineraries and how-to guides (pulling from information that already exists from every corner of the Internet). 

But ChatGPT can also kid around. And apparently, it can go a little too far. So far, in fact, that it's making 9/11 jokes.

Software developer @0xJim posted on X/Twitter his interaction with ChatGPT, starting with the prompt: “Why was 6 afraid of 7?” and getting the classic punchline “Because 7 8 (ate) 9!” 

Then, things got spicier when the user followed-up his original query with: “And why was 10 afraid?” ChatGPT’s response? “Because it was in the middle of 9/11.”

Apparently, other users followed suit and tested out ChatGPT's dark humor limits — with mixed results. 

Per one users’ ChatGPT script with the same prompts, the bot’s logic for the punchline was that the joke “plays on the idea that 10 is surrounded by 9 and 11, two numbers that might seem intimidating in this playful context.” But, the bot also seems to understand that it is, in fact, a double entendre that directly refers to the fall of the Twin Towers in 2001. “Sometimes even jokes can have a bit of a dark side,” it explained to one user. “Glad it gave you a laugh, though!”  

Neither OpenAI or Microsoft have released a statement regarding its prankster product, but it appears that such jokes are clearly within the limits of ChatGPT’s content filters and restrictions.