Thought experiment 1 – Humour and AI

by admin

In the Guardian article referenced below in Lively Stuff, scientist Richard Dawkins in self-deprecatory mode references a cartoon showing a chap hard at work late at night on his computer, while his wife’s voice from off begs him to come to bed. He cannot be disturbed, he cries, as “Someone is wrong on the Internet”.

In contemplation of such a scenario we will often talk of resonances, whole new meanings and something that is funny on so many levels. The hinterland of funny is a wondrous place and not the least of the wonders is how an AI of any stripe – narrow or deep, vanilla or chocolate – might fare within its borders. Chops can be licked in anticipation of a dinner table riff of the “What’s so funny about that?” variety, if the intelligences involved were human – or at least humans of the ilk of Stephen Fry with Amy Schumer, say; or Dawkins himself on tilting at windmills with an arch wit-jammer like Robin Williams, were he still with us, or Billy Connolly.

The sort of conversations that ensue when digital assistants and chatbots are set loose on YouTube still fall short of being interesting, much less funny, and much less than funny and interesting about what makes something funny and interesting. The ecstasy in contemplating the emerging whole that is greater than the sum of the parts lies at the heart of what it means to be an engaged human intelligence; and one challenge for the engineers of machine intelligence will lie in moving beyond the quantitative consolidations of every thing that has ever been uttered, were that even feasible. The day may be approaching when a chatbot might explain a joke, but minting one doesn’t yet appear on any horizon in our universe.