Existing chatbots are missing one key ingredient, Bitext explains in their blog post, “Linguistics to Create a Human-Like Chatbot.” Writer Clara García outlines the three linguistic factors an AI requires for human-like conversation—syntax, semantics, and pragmatics. The first two have been covered, but algorithms still struggle with pragmatics; that is, context and cultural knowledge. García illustrates:
Without pragmatics, our bot will never sound like a human, and what’s more important, it will not understand the user when she talks like one. If the user uses an idiom, cracks a joke, or uses the word ‘it’ referring to the skirt she was trying to buy, the bot will not understand her.
Before chatbots achieve that level of nuance, machine learning must improve. For more on how linguistics can help train smarter bots, download Bitext’s Deep Linguistic Analysis Benchmark.
Cynthia Murrell, April 27, 2017