Picture a nightmarish future in which half-witted “conversational user interfaces” drive us insane with mindless conversation and misunderstandings. As U.S. tech companies rush to emulate the success of China’s messaging platforms, this could be where we’re headed.
Dan Grover, a product manager for China’s wildly successful WeChat messaging service, argues that the chatbots and chatbot tools being developed by U.S. companies, such as Microsoft and Facebook, are inspired, in large part, by the many chat-based services now available in China. These chat services can be used for all sorts of tasks, including transferring money, paying restaurant bills, and looking up flight information.
Advertisement
The chatbots being developed in the U.S. are designed to perform tasks, such as searching for a flight or ordering a pizza, via a friendly back and forth with users. But, as Grover notes, China’s most successful chat interfaces forgo natural language in favor of more conventional input mechanisms such as multiple choice answers or buttons that appear in chat bubbles.
This story is only available to subscribers.
Don’t settle for half the story.
Get paywall-free access to technology news for the here and now.
That might seem backward, but it’s actually a more efficient way to order a pizza through a messaging platform than using a chatbot. Grover points out that conversational user interfaces require far more actions from a user than a simple messaging interface (73 taps compared to 16 when ordering a pizza, for example).
Nearly 700 million people use WeChat to do everything from pay bills to chat with friends.
To some extent, I fear, he is right that chatbots may be incredibly annoying. And things may be even more annoying if app designers fail to appreciate the significant challenges that remain with using computers to parse and respond to natural language.
But it would also be a mistake to give up on conversational interfaces altogether. For more open-ended tasks, some sort of dialogue may well be more desirable. And as devices like Amazon’s voice-controlled Echo device show, when speaking to a device you really need something that can converse, even if only on a very basic level.
