Could a computer be as good at sensing sarcasm in 140 characters as a person? If you think yes, the Secret Service might like to have a chat.
Next Gov picked up a work order the agency had put out, sketching out a full range of requested social network software capabilities, from the ability to work with Internet Explorer 8 to being able to comb through old tweets. Intriguingly, there was also this requested feature, “Ability to detect sarcasm and false positives.”
Of course, developing tech that can interpret and use language closer and closer to the way we ourselves do it is not new. Still, sarcasm — or any kind of speech where the literal meaning is layered with other (sometimes contradictory) meanings) — is a fairly sophisticated linguistic puzzle to unwrap, both for a computer, and sometimes even for people as well.
What do you think? How far away are we from software that really understands and interprets language as well as we can? And what would a world where that’s available look like?
Image: alphaspirit / Shutterstock.
