If language is bound up in living, if it is an expression of both sense and sensibility, then computers, being non-living, having no sensibility, will have a very difficult time mastering “natural-language processing” beyond a certain rudimentary level. The best solution, if you have a need to get computers to “understand” human communication, may to be avoid the problem altogether. Instead of figuring out how to get computers to understand natural language, you get people to speak artificial language, the language of computers. A good way to start is to encourage people to express themselves not through messy assemblages of fuzzily defined words but through neat, formal symbols — emoticons or emoji, for instance. When we speak with emoji, we’re speaking a language that machines can understand.
People like Mark Zuckerberg have always been uncomfortable with natural language. Now, they can do something about it.
I think we should be very concerned about this move by Facebook. In these contexts, I often think of a shrewd and troubling comment by Jaron Lanier: “The Turing test cuts both ways. You can't tell if a machine has gotten smarter or if you've just lowered your own standards of intelligence to such a degree that the machine seems smart. If you can have a conversation with a simulated person presented by an AI program, can you tell how far you've let your sense of personhood degrade in order to make the illusion work for you?” In this sense, the degradation of personhood is one of Facebook's explicit goals, and Facebook will increasingly require its users to cooperate in lowering their standards of intelligence and personhood.