I would be happy if I never had to type on a keyboard again.
It’s not possible today by a long shot, but there are signs that this future scenario could take place. I could have never imagined using my voice to edit an image in Photoshop, for example, but even that might be possible soon, judging from this Adobe Research proof of concept.
We’re now entering a new age, however. Alexa is everywhere. Siri is the go-to voicebot for online searches and directions. Google Home can carry on a conversation with you. The next steps in the “voice-first” revolution will be to go far beyond these basic functions.
For example, I’m hoping Microsoft makes radical improvements to Cortana soon. I want to browse websites by voice, have Cortana read articles to me, and talk to the bot about adjusting things like my screen resolution, advanced trackpad click options, going into battery saving mode, and much more. I’ve attempted to make Cortana part of my day on the desktop many times; it usually ends with me realizing the bot is primarily designed to set reminders and tell jokes.
On my phone, I want that same level of control, but Siri today is still a letdown. If I had a dollar for every time Siri merely presented web search results I’d have about $250.
And yet there is hope. Voice-first trumps textual interfaces in my digital world because of how much I’ve typed over the past 30 years and how frustrating and repetitive it can be. This is not a ding against chatbots, because there is definitely a big place in my digital world for texting when it’s impossible to talk to a bot — in an airplane, in a meeting, at a coffee shop.
Voice-first is a statement about priority. If I’m at my desk, I don’t pick up my phone first and type a question. I’ll say “Alexa, when do the Golden State Warriors play?” and get the answer. It has become a major habit for me now, something that is intertwined into my workflow. It’s not a stretch to think I’ll also ask Alexa to let me dictate a document soon. I’m also issuing voice commands for my garage door openers, my front door lock, the sprinklers in my garden (in the spring and summer), turning on the heat in the car I’m testing right now, and much more.
After dictation and smart home control? I want a bot that handles all of my email. I want to say things like “summarize all of my most important emails one by one” and have that work. I want to say “archive the emails I haven’t opened or read since last week” and then dictate a new message. This is possible today using bots like Cortana (using a Microsoft email app), but it is definitely not easy or efficient. For voice control to become standard and efficient, it has to work reliably every time.
In the next few years, voice control will go much further. Texting, making a phone call, troubleshooting a wireless problem, showing me pictures from my vacation, researching a complex topic — they are all voice-enabled today to some extent, but not exactly voice-friendly yet.
This is where true machine learning comes into play. For voicebots to revolutionize how we work and play, they will need to understand how we work and play. A bot needs to know I always text my college-age daughter (because she’s always studying or in class). If I tell a bot to “lock the doors” at night, it should be able to figure out that I mean my house and the car — because I’m going to bed. These are shortcuts and aids to make life and work easier without the fuss. I should be able to say “start a new document” to any voicebot and have that work perfectly. (And the bot should know I use Google Docs.) The bot should save an extra copy and download it as a Word file to whatever device I’m using. Maybe it should also start the email to send the doc.
And then, while this is more of a far-future concept, I want the voicebot to keep track of my entire day — not just let me dictate the document, but keep track of the document and even remind me that the document exists, that it needs my attention, or that it can be filed away. This is the companion voicebot we all recognize from movies like Her, but do not have yet in real life.
Developers, make it happen! I’m getting really sick of having to type everything.
This article was written by John Brandon from VentureBeat and was legally licensed through the NewsCred publisher network.