Artificial intelligence (AI) has given us the likes of Siri, Google and Amazon’s Alexa, but is talking to devices the future of the user interface (UI)?
I am sure many of you have asked silly questions when using Siri or Google, from “Knock, knock…who’s there?” to “What is the meaning of life?” But how many of you talk to them day in and day out? I suspect, like many of us, you often revert to typing your questions into the browser.
It might just be me getting old; because I tried it and talking out loud still feels uncomfortable — except maybe for in the privacy of my car. And I am not alone. In an article, Facebook founder Mark Zuckerberg, talking about his home project named Jarvis, said he wanted to be able give commands in and around his house to make life easier. However, he found out that in many cases text was more desirable, especially when there were other people around.
We will be confronted with the privacy issue everywhere we go, in public spaces and at home, because calling out commands will disturb strangers and family alike. Also, I think that texting and chatting gives less room for misinterpretation. After all, voice is just another layer on top of the text. What I mean by that is that none of the AIs today interpret commands directly; every sentence is first translated to text and then input into a computer for further processing (and vice-versa when it speaks to you).
In short, when the excitement over AI and voice UI winds down, my prediction is that the attention from speech will shift to text-based conversations. So, am I negative about the future of AI for human interaction? No, absolutely not. Although AI for natural language still has a long way to go, the latest generation of Natural Language Intelligence (NLI) is already very impressive, and the tools to train them get better and better by the day it seems. There are some NLIs out there that can carry on a good conversation on specific topics.
Above all, NLI’s will help us to solve the interaction hurdle with Internet of Things (IoT) devices. Handling Things by creating rules through anything other than an AI is a dead end. Many Things are used in so many different contexts that trying to interact through traditional rule setting is just too cumbersome. I wrote about this some time ago in this blog. NLI’s will replace many user interfaces, but it is us humans who will decide if we want to speak or write to them, based upon the context.
Finally, a word of warning to any company that thinks it doesn’t need to invest in this technology. I advise you to get familiar with and develop your own NLI. Because, if one day AI does become the new UI, the Alexas of this world will take you one more level away from your customers.