I have to admit, I found this to be a hard topic since it cut a fine line between rigorous abstract logic and more loose metaphysics. Most in attendance received the talk well, perhaps I will do another in the future.
A friend forwarded me a math riddle over the weekend :
Imagine: You and an artificial intelligence are to be subjected to a test. Each of you will anonymously submit a single English word to a (human) judge, who will try to determine which response came from a human being (as in a Turing test). Whoever is judged a human shall live; whoever is judged a machine shall be destroyed.
What word would you choose?
The trick, I think, is to realize that you are trying to psyche out the human judge – not the computer rival. This means finding a word that is subtly evocative of a human experience that would be hard for a computer to guess.
In other words, it would have to be a word associated with a human experience that nonetheless is not widely talked about in literature, history, or other media that a highly intelligent AI could search through. If we wanted to use a Venn Diagram we could express the problem like this:
This is not an easy task since, to be effective, the word would have to be on the edge of humanity’s ability to effectively express its own nature.
Any suggestions? What do you think are the most human words in our language?