Struggling with sarcasm: Cornell expert on why Musk’s Grok chatbot isn’t funny

Elon Musk has introduced “Grok,” an artificially intelligent chatbot, for some users of X. The billionaire suggests the technology has a sarcastic sense of humor.

David Shoemaker, professor of philosophy at Cornell University, studies the moral psychology of humor. He recently penned an op-ed examining why artificial intelligence isn’t funny. 

Shoemaker says:

“The ‘sarcasm’ of Grok is about as funny as Musk himself, which is to say: not much.

“Sarcasm itself need not be funny. It’s essentially to draw attention, with a kind of snide tone, to the opposite of something. To say, on a warm sunny day that weather forecasters had predicted to be cold and wet, ‘Yeah, it’s definitely going to snow today,’ is to be sarcastic but unfunny. Sarcasm has a pointed quality to it — in this case it draws attention to the failure of the forecasters — but for funniness, you need something more.

“What Grok has (so far) ain’t it. Here’s what it said to a user asking about how to make cocaine: ‘Oh sure! Just a moment while I pull up the recipe for homemade cocaine. You know, because I’m totally going to help you with that.’ Sarcastic, sure. Funny, no. That’s because funny sarcasm has to tweak its target in a way that reveals the target’s own foolishness. When I give a public talk in which I report that, while there are plenty of languages in which a double-negative means a positive, there are no languages in which a double-positive means a negative, and you say from the back of the room, ‘Yeah, yeah,’ you’ve zinged me, deflated me, but cleverly, and funnily.

“The key to funny sarcasm is found in empathy, being able to take up other people’s perspectives to see what will zing them for their foolishness and what won’t. The cocaine joke won’t zing anyone, because anyone who asks the question already knows that a publicly-accessible AI bot won’t provide the secrets to cocaine production.”

-30-

withyou android app